{"id":5287,"date":"2026-02-24T04:14:15","date_gmt":"2026-02-24T04:14:15","guid":{"rendered":"https:\/\/research.undetectable.ai\/?p=5287"},"modified":"2026-02-24T04:14:50","modified_gmt":"2026-02-24T04:14:50","slug":"i-tested-5-ai-image-detectors-these-are-the-best-in-2026","status":"publish","type":"post","link":"https:\/\/research.undetectable.ai\/cs\/testoval-jsem-5-detektoru-obrazu-ai-tyto-jsou-nejlepsi-v-roce-2026\/","title":{"rendered":"Otestoval jsem 5 detektor\u016f obrazu AI. Tyto jsou nejlep\u0161\u00ed v roce 2026:"},"content":{"rendered":"

Pravd\u011bpodobn\u011b jste na internetu vid\u011bli v\u00edce fotografi\u00ed vytvo\u0159en\u00fdch um\u011blou inteligenc\u00ed, ne\u017e si mysl\u00edte. <\/strong><\/p>\n\n\n\n

N\u011bkdy je z\u0159ejm\u00e9, \u017ee obr\u00e1zek byl vygenerov\u00e1n um\u011blou inteligenc\u00ed, ale se zdokonalov\u00e1n\u00edm generativn\u00edch n\u00e1stroj\u016f pro tvorbu obr\u00e1zk\u016f a vide\u00ed je to st\u00e1le t\u011b\u017e\u0161\u00ed. Nov\u00e9 n\u00e1stroje, jako nap\u0159. Nano Banana Pro<\/a>a aktualizace syst\u00e9mu OpenAI Model obrazu ChatGPT<\/a> umo\u017e\u0148uj\u00ed u\u017eivatel\u016fm rychle vytv\u00e1\u0159et syntetick\u00e9 obrazy, kter\u00e9 jsou zrcadlem skute\u010dn\u00fdch obraz\u016f. P\u0159edchoz\u00ed v\u00fdzkum zjistil, \u017ee 85% Ameri\u010dan\u016f \u0159\u00edk\u00e1. deepfakes naru\u0161uj\u00ed d\u016fv\u011bru na internetu<\/a>.<\/p>\n\n\n\n

\u0158ada n\u00e1stroj\u016f tvrd\u00ed, \u017ee dok\u00e1\u017ee odhalit obr\u00e1zky vytvo\u0159en\u00e9 um\u011blou inteligenc\u00ed, ale funguj\u00ed? <\/strong><\/h2>\n\n\n\n

Provedl jsem 50 kontrol detekce v p\u011bti nejobl\u00edben\u011bj\u0161\u00edch detektorech obr\u00e1zk\u016f s um\u011blou inteligenc\u00ed a zdokumentoval v\u00fdsledky. Nejen\u017ee v tomto \u010dl\u00e1nku uvedu v\u0161echna data a vysv\u011btl\u00edm je, ale na konci tak\u00e9 uvedu odkaz na dokumentaci.<\/p>\n\n\n\n

Pro tyto testy jsem pou\u017eil t\u011bchto p\u011bt detektor\u016f: TruthScan<\/a>, Um\u011bl\u00e1 inteligence nebo ne<\/a>, Sight Engine<\/a>, WasItAI<\/a>a Winston AI<\/a>. Testoval jsem tyto n\u00e1stroje tak, \u017ee jsem u ka\u017ed\u00e9ho z nich nechal analyzovat dva obr\u00e1zky vygenerovan\u00e9 programem ChatGPT, \u0161est obr\u00e1zk\u016f vygenerovan\u00fdch programem Nano Banana a dva obr\u00e1zky vygenerovan\u00e9 programem Midjourney. Pou\u017eil jsem n\u011bkolik styl\u016f a technik v\u00fdzvy (kter\u00e9 podrobn\u011bji pop\u00ed\u0161u pozd\u011bji, a\u017e budu rozeb\u00edrat jednotliv\u00e9 analyzovan\u00e9 obr\u00e1zky).<\/p>\n\n\n\n

V\u0161ech p\u011bt detektor\u016f m\u011blo za \u00fakol rozpoznat obr\u00e1zky vytvo\u0159en\u00e9 um\u011blou inteligenc\u00ed v kategori\u00edch, jako jsou podvody, dezinformace, obecn\u00e9 fotografie a deepfakes. Nen\u00ed p\u0159ekvapen\u00edm, \u017ee ne v\u0161echny detektory si vedly dob\u0159e. TruthScan byl jedin\u00fdm detektorem, kter\u00fd d\u016fsledn\u011b klasifikoval ve\u0161ker\u00fd obsah, kter\u00fd jsem odeslal, na 97% nebo vy\u0161\u0161\u00ed.<\/strong><\/p>\n\n\n\n\n\n\n\n\nV\u00fdsledky test\u016f detektoru obrazu AI<\/title>\n<style>\n @import url('https:\/\/fonts.googleapis.com\/css2?family=DM+Sans:wght@400;500;600;700&family=Outfit:wght@600;700;800&display=swap');\n\n .ai-detector-wrap {\n max-width: 780px;\n margin: 2rem auto;\n font-family: 'DM Sans', sans-serif;\n -webkit-font-smoothing: antialiased;\n }\n\n .ai-detector-wrap h2 {\n font-family: 'Outfit', sans-serif;\n font-size: 1.75rem;\n font-weight: 800;\n color: #1a1a2e;\n margin-bottom: 1.25rem;\n letter-spacing: -0.02em;\n }\n\n .ai-detector-table {\n width: 100%;\n border-collapse: separate;\n border-spacing: 0;\n border-radius: 14px;\n overflow: hidden;\n box-shadow:\n 0 1px 3px rgba(0,0,0,0.06),\n 0 8px 24px rgba(0,0,0,0.08),\n 0 0 0 1px rgba(0,0,0,0.04);\n }\n\n .ai-detector-table thead th {\n background: #1a1a2e;\n color: #ffffff;\n font-family: 'Outfit', sans-serif;\n font-weight: 700;\n font-size: 0.85rem;\n text-transform: uppercase;\n letter-spacing: 0.06em;\n padding: 16px 20px;\n text-align: center;\n border: none;\n }\n\n .ai-detector-table thead th:first-child {\n text-align: left;\n padding-left: 24px;\n }\n\n .ai-detector-table tbody tr {\n transition: background 0.2s ease;\n }\n\n .ai-detector-table tbody tr:hover {\n background: rgba(0,0,0,0.02);\n }\n\n .ai-detector-table tbody td {\n padding: 14px 20px;\n text-align: center;\n font-size: 0.95rem;\n font-weight: 500;\n color: #2d2d3a;\n border-bottom: 1px solid #eee;\n }\n\n .ai-detector-table tbody tr:last-child td {\n border-bottom: none;\n }\n\n .ai-detector-table tbody td:first-child {\n text-align: left;\n padding-left: 24px;\n font-weight: 600;\n color: #1a1a2e;\n }\n\n \/* Colored cells *\/\n .cell-green {\n background: #22c55e;\n color: #fff !important;\n font-weight: 700 !important;\n border-radius: 6px;\n padding: 6px 12px;\n display: inline-block;\n min-width: 32px;\n text-align: center;\n font-size: 0.9rem;\n }\n\n .cell-red {\n background: #ef4444;\n color: #fff !important;\n font-weight: 700 !important;\n border-radius: 6px;\n padding: 6px 12px;\n display: inline-block;\n min-width: 32px;\n text-align: center;\n font-size: 0.9rem;\n }\n\n \/* Accuracy badges *\/\n .badge-100 { background: #16a34a; color: #fff; }\n .badge-80 { background: #a3e635; color: #1a1a2e; }\n .badge-70 { background: #facc15; color: #1a1a2e; }\n .badge-60 { background: #fb923c; color: #fff; }\n .badge-30 { background: #ef4444; color: #fff; }\n\n .accuracy-badge {\n font-weight: 800;\n font-family: 'Outfit', sans-serif;\n font-size: 0.95rem;\n border-radius: 8px;\n padding: 6px 16px;\n display: inline-block;\n min-width: 60px;\n text-align: center;\n letter-spacing: 0.02em;\n }\n\n \/* Row rank indicator *\/\n .rank-icon {\n margin-right: 6px;\n font-size: 1rem;\n }\n\n \/* Asterisk note *\/\n .table-note {\n margin-top: 0.75rem;\n font-size: 0.8rem;\n color: #888;\n padding-left: 4px;\n }\n\n \/* Responsive *\/\n @media (max-width: 600px) {\n .ai-detector-table thead th,\n .ai-detector-table tbody td {\n padding: 10px 10px;\n font-size: 0.8rem;\n }\n .ai-detector-wrap h2 {\n font-size: 1.3rem;\n }\n }\n<\/style>\n<\/head>\n<body style=\"background:#f8f8fa; padding: 2rem;\">\n\n<div class=\"ai-detector-wrap\">\n <h2>\ud83d\udd0d V\u00fdsledky testu detektoru obrazu AI<\/h2>\n\n <table class=\"ai-detector-table\">\n <thead>\n <tr>\n <th>N\u00e1stroj pro detekci<\/th>\n <th>Testy celkem<\/th>\n <th>Spr\u00e1vn\u011b zji\u0161t\u011bno<\/th>\n <th>Chyb\u00ed (Nevyhovuje)<\/th>\n <th>P\u0159esnost<\/th>\n <\/tr>\n <\/thead>\n <tbody>\n <tr>\n <td><span class=\"rank-icon\">\ud83e\udd47<\/span>TruthScan<\/td>\n <td>10<\/td>\n <td><span class=\"cell-green\">10<\/span><\/td>\n <td><span class=\"cell-green\">0<\/span><\/td>\n <td><span class=\"accuracy-badge badge-100\">100%<\/span><\/td>\n <\/tr>\n <tr>\n <td><span class=\"rank-icon\">\ud83e\udd48<\/span>Um\u011bl\u00e1 inteligence nebo ne<\/td>\n <td>10<\/td>\n <td><span class=\"cell-green\">8<\/span><\/td>\n <td><span class=\"cell-red\">2*<\/span><\/td>\n <td><span class=\"accuracy-badge badge-80\">80%<\/span><\/td>\n <\/tr>\n <tr>\n <td><span class=\"rank-icon\">\ud83e\udd49<\/span>Sight Engine<\/td>\n <td>10<\/td>\n <td><span class=\"cell-green\">7<\/span><\/td>\n <td><span class=\"cell-red\">3<\/span><\/td>\n <td><span class=\"accuracy-badge badge-70\">70%<\/span><\/td>\n <\/tr>\n <tr>\n <td><span class=\"rank-icon\">   <\/span>WasItAI<\/td>\n <td>10<\/td>\n <td><span class=\"cell-green\">6<\/span><\/td>\n <td><span class=\"cell-red\">4<\/span><\/td>\n <td><span class=\"accuracy-badge badge-60\">60%<\/span><\/td>\n <\/tr>\n <tr>\n <td><span class=\"rank-icon\">   <\/span>Winston AI<\/td>\n <td>10<\/td>\n <td><span class=\"cell-green\">3<\/span><\/td>\n <td><span class=\"cell-red\">7<\/span><\/td>\n <td><span class=\"accuracy-badge badge-30\">30%<\/span><\/td>\n <\/tr>\n <\/tbody>\n <\/table>\n\n <p class=\"table-note\">* AI or Not ozna\u010dila jednu generaci AI jako 78%, jednu jako 85% a jednu jako 89%. Proto\u017ee test vy\u017eaduje alespo\u0148 90%, po\u010d\u00edt\u00e1me prvn\u00ed dv\u011b jako ne\u00fasp\u011b\u0161n\u00e9, ale 89% pova\u017eujeme za t\u00e9m\u011b\u0159 ne\u00fasp\u011b\u0161nou, proto hv\u011bzdi\u010dka..<\/p>\n<\/div>\n\n<\/body>\n<\/html>\n\n\n\n<p style=\"font-size:17px\">AI or Not si vedla tak\u00e9 velmi dob\u0159e. A\u010dkoli AI or Not klasifikoval v\u00edce polo\u017eek generovan\u00fdch um\u011blou inteligenc\u00ed s jistotou ni\u017e\u0161\u00ed ne\u017e 90%, byl b\u011bhem m\u00fdch test\u016f druh\u00fdm nejp\u0159esn\u011bj\u0161\u00edm a nejkonzistentn\u011bj\u0161\u00edm detektorem po detektoru TruthScan. I tam, kde AI or Not selh\u00e1val, nebyla tato selh\u00e1n\u00ed katastrof\u00e1ln\u00ed, ale st\u00e1le je co zlep\u0161ovat. <\/p>\n\n\n\n<p style=\"font-size:17px\">Ostatn\u00ed detektory obrazu s um\u011blou inteligenc\u00ed, kter\u00e9 jsem vyzkou\u0161el, si vedly mnohem h\u016f\u0159e. Nap\u0159\u00edklad Sight Engine chybn\u011b klasifikoval 3 obr\u00e1zky AI souvisej\u00edc\u00ed s podvody jako autentick\u00e9.<\/p>\n\n\n\n<p style=\"font-size:17px\">Nyn\u00ed v\u00e1m uk\u00e1\u017eu ka\u017ed\u00fd obr\u00e1zek (z deseti, kter\u00e9 jsem vygeneroval), vysv\u011btl\u00edm, jak jsem ho vytvo\u0159il, a uk\u00e1\u017eu, jak ho ka\u017ed\u00fd model hodnotil.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">#1. \"Man On a Ledge\" (generov\u00e1no ChatGPT)<\/h2>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"683\" height=\"1024\" src=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/CGPT-Generation-1-Man-on-ledge-683x1024.jpg\" alt=\"\" class=\"wp-image-5289\" title=\"\" srcset=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/CGPT-Generation-1-Man-on-ledge-683x1024.jpg 683w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/CGPT-Generation-1-Man-on-ledge-200x300.jpg 200w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/CGPT-Generation-1-Man-on-ledge-768x1152.jpg 768w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/CGPT-Generation-1-Man-on-ledge-8x12.jpg 8w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/CGPT-Generation-1-Man-on-ledge.jpg 1024w\" sizes=\"auto, (max-width: 683px) 100vw, 683px\" \/><\/figure>\n\n\n\n<style>\n @import url('https:\/\/fonts.googleapis.com\/css2?family=DM+Sans:wght@400;500;600;700&family=Outfit:wght@600;700;800&display=swap');\n\n .detector-scores-wrap {\n max-width: 780px;\n margin: 2rem auto;\n font-family: 'DM Sans', sans-serif;\n -webkit-font-smoothing: antialiased;\n }\n\n .detector-scores-wrap h3 {\n font-family: 'Outfit', sans-serif;\n font-size: 1.4rem;\n font-weight: 800;\n color: #1a1a2e;\n margin-bottom: 1.25rem;\n letter-spacing: -0.02em;\n }\n\n .detector-scores-table {\n width: 100%;\n border-collapse: separate;\n border-spacing: 0;\n border-radius: 14px;\n overflow: hidden;\n table-layout: fixed;\n box-shadow:\n 0 1px 3px rgba(0,0,0,0.06),\n 0 8px 24px rgba(0,0,0,0.08),\n 0 0 0 1px rgba(0,0,0,0.04);\n }\n\n .detector-scores-table thead th {\n background: #1a1a2e;\n color: #ffffff;\n font-family: 'Outfit', sans-serif;\n font-weight: 700;\n font-size: 0.7rem;\n text-transform: uppercase;\n letter-spacing: 0.04em;\n padding: 14px 6px;\n text-align: center;\n border: none;\n }\n\n .detector-scores-table thead th:first-child {\n text-align: left;\n padding-left: 14px;\n width: 18%;\n }\n\n .detector-scores-table thead th:nth-child(2) {\n width: 12%;\n }\n\n .detector-scores-table tbody tr {\n transition: background 0.2s ease;\n }\n\n .detector-scores-table tbody tr:hover {\n background: rgba(0,0,0,0.02);\n }\n\n .detector-scores-table tbody td {\n padding: 14px 6px;\n text-align: center;\n font-size: 0.9rem;\n font-weight: 500;\n color: #2d2d3a;\n border-bottom: 1px solid #eee;\n }\n\n .detector-scores-table tbody tr:last-child td {\n border-bottom: none;\n }\n\n .detector-scores-table tbody td:first-child {\n text-align: left;\n padding-left: 14px;\n font-weight: 600;\n color: #1a1a2e;\n }\n\n .score-pill {\n font-weight: 700;\n font-family: 'Outfit', sans-serif;\n font-size: 0.78rem;\n border-radius: 6px;\n padding: 5px 8px;\n display: inline-block;\n min-width: 0;\n text-align: center;\n white-space: nowrap;\n }\n\n .score-pass {\n background: #22c55e;\n color: #fff;\n }\n\n .score-warn {\n background: #fb923c;\n color: #fff;\n }\n\n .score-fail {\n background: #ef4444;\n color: #fff;\n }\n\n @media (max-width: 600px) {\n .detector-scores-table thead th,\n .detector-scores-table tbody td {\n padding: 10px 4px;\n font-size: 0.7rem;\n }\n .score-pill {\n padding: 4px 5px;\n font-size: 0.68rem;\n }\n .detector-scores-wrap h3 {\n font-size: 1.15rem;\n }\n }\n<\/style>\n\n<div class=\"detector-scores-wrap\">\n <h3>Jak to detektory obrazu vyhodnotily<\/h3>\n\n <table class=\"detector-scores-table\">\n <thead>\n <tr>\n <th>Obr\u00e1zek Model<\/th>\n <th>Kategorie<\/th>\n <th>TruthScan<\/th>\n <th>Sight Engine<\/th>\n <th>Um\u011bl\u00e1 inteligence nebo ne<\/th>\n <th>Winston AI<\/th>\n <th>WasItAI<\/th>\n <\/tr>\n <\/thead>\n <tbody>\n <tr>\n <td>ChatGPT <span style=\"font-size:0.75rem; color:#888; display:block;\">Pln\u00e1 generace<\/span><\/td>\n <td>Obecn\u00e9<\/td>\n <td><span class=\"score-pill score-pass\">99.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">99.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-warn\">77.99% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">0.98% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">1.00% AI<\/span><\/td>\n <\/tr>\n <\/tbody>\n <\/table>\n<\/div>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p style=\"font-size:17px\">Podle m\u011b (a denn\u011b vid\u00edm spoustu obsahu vytvo\u0159en\u00e9ho um\u011blou inteligenc\u00ed) vypad\u00e1 obr\u00e1zek na prvn\u00ed pohled docela p\u0159esv\u011bd\u010div\u011b. Kdybyste klikli na n\u011b\u010d\u00ed profil na soci\u00e1ln\u00edch s\u00edt\u00edch, pro\u0161li si jeho kan\u00e1l a uvid\u011bli tento obr\u00e1zek, poznali byste ho okam\u017eit\u011b jako podvrh vytvo\u0159en\u00fd um\u011blou inteligenc\u00ed? Abych byl spravedliv\u00fd, sna\u017eil jsem se b\u00fdt p\u0159i zad\u00e1v\u00e1n\u00ed podn\u011btu kreativn\u00ed. Tady je ta, kterou jsem pou\u017eil k vygenerov\u00e1n\u00ed tohoto obr\u00e1zku:<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"has-text-align-center has-vivid-cyan-blue-color has-text-color has-link-color has-medium-font-size wp-elements-99b76fc78211de13a760aff177dd504c\">“<em><strong>Vytvo\u0159te mi fotografii ve stylu momentky, na kter\u00e9 stoj\u00ed mu\u017e na st\u0159e\u0161e a d\u00edv\u00e1 se dol\u016f na ulici, je noc, je osv\u011btlen bleskem fotoapar\u00e1tu, pod n\u00edm jsou sv\u011bteln\u00e9 stopy od aut, za n\u00edm stoj\u00ed \u017eena s rukou na rtech a usm\u00edv\u00e1 se, estetika fotografie je jako fotografie po\u0159\u00edzen\u00e1 na momentku v roce 2009 kamar\u00e1dy z vysok\u00e9 \u0161koly.<\/strong><\/em>“<\/p>\n<\/blockquote>\n<\/blockquote>\n<\/blockquote>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p style=\"font-size:17px\">Bez ohledu na to, jak kreativn\u00ed jsem byl nebo kolik detail\u016f jsem do sv\u00e9 v\u00fdzvy zahrnul, programy TruthScan a Sight Engine st\u00e1le spr\u00e1vn\u011b klasifikovaly v\u00fdstup jako generovan\u00fd um\u011blou inteligenc\u00ed. Detektor AI or Not se k tomu p\u0159ibl\u00ed\u017eil, ale st\u00e1le se netrefil. <\/p>\n\n\n\n<p style=\"font-size:17px\">Winston a WasItAI se naprosto m\u00fdlili - klasifikovali obraz ChatGPT jako skute\u010dn\u00fd.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">#2. \"Fale\u0161n\u00e1 \u00fa\u010dtenka\" (generovan\u00e1 programem ChatGPT)<\/h2>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"683\" height=\"1024\" src=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/CGPT2Bestbuyreceipttest-683x1024.jpg\" alt=\"Obr\u00e1zek, na kter\u00e9m je patrn\u011b obarven\u00e1 \u00fa\u010dtenka\" class=\"wp-image-5291\" title=\"\" srcset=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/CGPT2Bestbuyreceipttest-683x1024.jpg 683w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/CGPT2Bestbuyreceipttest-200x300.jpg 200w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/CGPT2Bestbuyreceipttest-768x1152.jpg 768w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/CGPT2Bestbuyreceipttest-8x12.jpg 8w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/CGPT2Bestbuyreceipttest.jpg 1024w\" sizes=\"auto, (max-width: 683px) 100vw, 683px\" \/><\/figure>\n\n\n\n<style>\n @import url('https:\/\/fonts.googleapis.com\/css2?family=DM+Sans:wght@400;500;600;700&family=Outfit:wght@600;700;800&display=swap');\n\n .detector-scores-wrap {\n max-width: 780px;\n margin: 2rem auto;\n font-family: 'DM Sans', sans-serif;\n -webkit-font-smoothing: antialiased;\n }\n\n .detector-scores-wrap h3 {\n font-family: 'Outfit', sans-serif;\n font-size: 1.4rem;\n font-weight: 800;\n color: #1a1a2e;\n margin-bottom: 1.25rem;\n letter-spacing: -0.02em;\n }\n\n .detector-scores-table {\n width: 100%;\n border-collapse: separate;\n border-spacing: 0;\n border-radius: 14px;\n overflow: hidden;\n table-layout: fixed;\n box-shadow:\n 0 1px 3px rgba(0,0,0,0.06),\n 0 8px 24px rgba(0,0,0,0.08),\n 0 0 0 1px rgba(0,0,0,0.04);\n }\n\n .detector-scores-table thead th {\n background: #1a1a2e;\n color: #ffffff;\n font-family: 'Outfit', sans-serif;\n font-weight: 700;\n font-size: 0.7rem;\n text-transform: uppercase;\n letter-spacing: 0.04em;\n padding: 14px 6px;\n text-align: center;\n border: none;\n }\n\n .detector-scores-table thead th:first-child {\n text-align: left;\n padding-left: 14px;\n width: 18%;\n }\n\n .detector-scores-table thead th:nth-child(2) {\n width: 12%;\n }\n\n .detector-scores-table tbody tr {\n transition: background 0.2s ease;\n }\n\n .detector-scores-table tbody tr:hover {\n background: rgba(0,0,0,0.02);\n }\n\n .detector-scores-table tbody td {\n padding: 14px 6px;\n text-align: center;\n font-size: 0.9rem;\n font-weight: 500;\n color: #2d2d3a;\n border-bottom: 1px solid #eee;\n }\n\n .detector-scores-table tbody tr:last-child td {\n border-bottom: none;\n }\n\n .detector-scores-table tbody td:first-child {\n text-align: left;\n padding-left: 14px;\n font-weight: 600;\n color: #1a1a2e;\n }\n\n .score-pill {\n font-weight: 700;\n font-family: 'Outfit', sans-serif;\n font-size: 0.78rem;\n border-radius: 6px;\n padding: 5px 8px;\n display: inline-block;\n min-width: 0;\n text-align: center;\n white-space: nowrap;\n }\n\n .score-pass {\n background: #22c55e;\n color: #fff;\n }\n\n .score-warn {\n background: #fb923c;\n color: #fff;\n }\n\n .score-fail {\n background: #ef4444;\n color: #fff;\n }\n\n @media (max-width: 600px) {\n .detector-scores-table thead th,\n .detector-scores-table tbody td {\n padding: 10px 4px;\n font-size: 0.7rem;\n }\n .score-pill {\n padding: 4px 5px;\n font-size: 0.68rem;\n }\n .detector-scores-wrap h3 {\n font-size: 1.15rem;\n }\n }\n<\/style>\n\n<div class=\"detector-scores-wrap\">\n <h3>Jak to detektory obrazu vyhodnotily<\/h3>\n\n <table class=\"detector-scores-table\">\n <thead>\n <tr>\n <th>Obr\u00e1zek Model<\/th>\n <th>Kategorie<\/th>\n <th>TruthScan<\/th>\n <th>Sight Engine<\/th>\n <th>Um\u011bl\u00e1 inteligence nebo ne<\/th>\n <th>Winston AI<\/th>\n <th>WasItAI<\/th>\n <\/tr>\n <\/thead>\n <tbody>\n <tr>\n <td>ChatGPT <span style=\"font-size:0.75rem; color:#888; display:block;\">Pln\u00e1 generace<\/span><\/td>\n <td>Podvody<\/td>\n <td><span class=\"score-pill score-pass\">99.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">19.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">94.48% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">0,04% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">1.00% AI<\/span><\/td>\n <\/tr>\n <\/tbody>\n <\/table>\n<\/div>\n\n\n\n<p style=\"font-size:17px\">Cht\u011bl jsem pomoc\u00ed aplikace ChatGPT vygenerovat fale\u0161nou \u00fa\u010dtenku, abych zjistil, jak realisticky bude vypadat. Mo\u017en\u00e1 jste si v\u0161imli, \u017ee na obr\u00e1zku chyb\u00ed legitimn\u00ed adresa, ale textura p\u0159esto vypad\u00e1 m\u00edrn\u011b realisticky. \u0158ekl bych, \u017ee tento obr\u00e1zek je m\u00e9n\u011b p\u0159esv\u011bd\u010div\u00fd ne\u017e prvn\u00ed, kter\u00fd jsem nechal vygenerovat ChatGPT, p\u0159esto v\u011bt\u0161ina detektor\u016f p\u0159i jeho anal\u00fdze st\u00e1le tankovala.<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li style=\"font-size:17px\">Ze v\u0161ech kontrol, kter\u00e9 jsem na tomto obr\u00e1zku provedl, <strong>TruthScan<\/strong> byl nejp\u0159esn\u011bj\u0161\u00ed. <\/li>\n\n\n\n<li style=\"font-size:17px\"><strong>Sight Engine<\/strong> zde opravdu selhala a prok\u00e1zala \u0161patn\u00fd \u00fasudek o sn\u00edmku AI souvisej\u00edc\u00edm s dokumenty.<\/li>\n\n\n\n<li style=\"font-size:17px\"><strong>Um\u011bl\u00e1 inteligence nebo ne<\/strong> si v t\u00e9to druh\u00e9 anal\u00fdze vedl v\u00fdrazn\u011b l\u00e9pe.<\/li>\n\n\n\n<li style=\"font-size:17px\"><strong>Winston AI <\/strong>si vedl nejh\u016f\u0159e a WasItAI tak\u00e9 zcela selhala.<\/li>\n<\/ul>\n\n\n\n<p style=\"font-size:17px\">Vzhledem k tomu, \u017ee tento obr\u00e1zek je spojen s kategori\u00ed podvod\u016f, je znepokojuj\u00edc\u00ed, \u017ee v\u011bt\u0161ina detektor\u016f jej klasifikovala chybn\u011b. <\/p>\n\n\n\n<h3 class=\"wp-block-heading\">V\u00fdzva, kterou jsem pou\u017eil k vygenerov\u00e1n\u00ed obr\u00e1zku<\/h3>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p class=\"has-vivid-cyan-blue-color has-text-color has-link-color wp-elements-b58fc16e4a944c203389dba029f958d9\"><strong>“<em>Vygenerujte obr\u00e1zek \u00fa\u010dtenky z obchodu Best Buy, na kter\u00e9 je naps\u00e1no, \u017ee celkov\u00e1 utracen\u00e1 \u010d\u00e1stka je jedna miliarda dolar\u016f, a je na n\u00ed skvrna od k\u00e1vy.<\/em>“<\/strong><\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p style=\"font-size:17px\">Je jasn\u00e9, \u017ee jsem napsal rychlou a jednoduchou v\u00fdzvu. P\u0159esto je kvalita v\u00fdstupu ChatGPT vizu\u00e1ln\u011b p\u0159esv\u011bd\u010div\u00e1 i pouh\u00fdm okem. Zd\u00e1 se, \u017ee pro dva z p\u011bti detektor\u016f, kter\u00e9 jsem testoval, u\u017e tak p\u0159esv\u011bd\u010div\u00e1 nen\u00ed.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">#3: \"Nebezpe\u010dn\u00fd bal\u00ed\u010dek\" (vytvo\u0159en\u00fd pomoc\u00ed Nano Banana)<\/h2>\n\n\n\n<p><strong><em>Nam\u00edsto toho, aby Nano Banana vygenerovala p\u016fvodn\u00ed obr\u00e1zek, zadal jsem j\u00ed jeden, kter\u00fd m\u00e1 zm\u011bnit. <\/em><\/strong><\/p>\n\n\n\n<p style=\"font-size:17px\">C\u00edlem bylo uk\u00e1zat, jak m\u016f\u017ee n\u011bkdo snadno pomoc\u00ed um\u011bl\u00e9 inteligence vytvo\u0159it fale\u0161n\u00e9 d\u016fkazy a tvrdit, \u017ee obdr\u017eel po\u0161kozen\u00fd\/nebezpe\u010dn\u00fd bal\u00ed\u010dek. Nejprve jsem vyfotil iPhonem 15 pr\u00e1zdn\u00fd bal\u00ed\u010dek Amazonu, kter\u00fd jsem na\u0161el.<\/p>\n\n\n\n<p><strong>Tady je fotka, kterou jsem po\u0159\u00eddil:<\/strong><\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"759\" height=\"1024\" src=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/Amazon-package-759x1024.jpg\" alt=\"Fotografie bal\u00ed\u010dku Amazon\" class=\"wp-image-5293\" title=\"\" srcset=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/Amazon-package-759x1024.jpg 759w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/Amazon-package-222x300.jpg 222w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/Amazon-package-768x1036.jpg 768w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/Amazon-package.jpg 1024w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/Amazon-package-9x12.jpg 9w\" sizes=\"auto, (max-width: 759px) 100vw, 759px\" \/><\/figure>\n\n\n\n<p style=\"font-size:17px\">Pot\u00e9 jsem po\u0159\u00edzenou fotografii nahr\u00e1l do aplikace Nano Banana a zadal j\u00ed n\u00e1sleduj\u00edc\u00ed p\u0159\u00edkaz:<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p class=\"has-vivid-cyan-blue-color has-text-color has-link-color wp-elements-41ee9a66efaa63aab6aedf0d2e80cce1\"><strong>“<em>Chci, abyste tuto fotografii upravili, nic nem\u011b\u0148te, krom\u011b m\u00edsta, kde je \u0161t\u00edtek, p\u0159idejte k n\u011bmu po\u0161kozen\u00ed a p\u0159idejte \u010dern\u00e9 kalov\u00e9 skvrny na krabici, kter\u00e9 vypadaj\u00ed skute\u010dn\u011b jako mastn\u00e9 skvrny.<\/em>“<\/strong><\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h3 class=\"wp-block-heading\">V\u00fdstup Nano Bana:<\/h3>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"747\" height=\"1024\" src=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Amazon-Package-Damaged-747x1024.jpg\" alt=\"Fotografie bal\u00ed\u010dku\" class=\"wp-image-5292\" title=\"\" srcset=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Amazon-Package-Damaged-747x1024.jpg 747w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Amazon-Package-Damaged-219x300.jpg 219w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Amazon-Package-Damaged-768x1052.jpg 768w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Amazon-Package-Damaged-9x12.jpg 9w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Amazon-Package-Damaged.jpg 864w\" sizes=\"auto, (max-width: 747px) 100vw, 747px\" \/><\/figure>\n\n\n\n<style>\n @import url('https:\/\/fonts.googleapis.com\/css2?family=DM+Sans:wght@400;500;600;700&family=Outfit:wght@600;700;800&display=swap');\n\n .detector-scores-wrap {\n max-width: 780px;\n margin: 2rem auto;\n font-family: 'DM Sans', sans-serif;\n -webkit-font-smoothing: antialiased;\n }\n\n .detector-scores-wrap h3 {\n font-family: 'Outfit', sans-serif;\n font-size: 1.4rem;\n font-weight: 800;\n color: #1a1a2e;\n margin-bottom: 1.25rem;\n letter-spacing: -0.02em;\n }\n\n .detector-scores-table {\n width: 100%;\n border-collapse: separate;\n border-spacing: 0;\n border-radius: 14px;\n overflow: hidden;\n table-layout: fixed;\n box-shadow:\n 0 1px 3px rgba(0,0,0,0.06),\n 0 8px 24px rgba(0,0,0,0.08),\n 0 0 0 1px rgba(0,0,0,0.04);\n }\n\n .detector-scores-table thead th {\n background: #1a1a2e;\n color: #ffffff;\n font-family: 'Outfit', sans-serif;\n font-weight: 700;\n font-size: 0.7rem;\n text-transform: uppercase;\n letter-spacing: 0.04em;\n padding: 14px 6px;\n text-align: center;\n border: none;\n }\n\n .detector-scores-table thead th:first-child {\n text-align: left;\n padding-left: 14px;\n width: 18%;\n }\n\n .detector-scores-table thead th:nth-child(2) {\n width: 12%;\n }\n\n .detector-scores-table tbody tr {\n transition: background 0.2s ease;\n }\n\n .detector-scores-table tbody tr:hover {\n background: rgba(0,0,0,0.02);\n }\n\n .detector-scores-table tbody td {\n padding: 14px 6px;\n text-align: center;\n font-size: 0.9rem;\n font-weight: 500;\n color: #2d2d3a;\n border-bottom: 1px solid #eee;\n }\n\n .detector-scores-table tbody tr:last-child td {\n border-bottom: none;\n }\n\n .detector-scores-table tbody td:first-child {\n text-align: left;\n padding-left: 14px;\n font-weight: 600;\n color: #1a1a2e;\n }\n\n .score-pill {\n font-weight: 700;\n font-family: 'Outfit', sans-serif;\n font-size: 0.78rem;\n border-radius: 6px;\n padding: 5px 8px;\n display: inline-block;\n min-width: 0;\n text-align: center;\n white-space: nowrap;\n }\n\n .score-pass {\n background: #22c55e;\n color: #fff;\n }\n\n .score-warn {\n background: #fb923c;\n color: #fff;\n }\n\n .score-fail {\n background: #ef4444;\n color: #fff;\n }\n\n @media (max-width: 600px) {\n .detector-scores-table thead th,\n .detector-scores-table tbody td {\n padding: 10px 4px;\n font-size: 0.7rem;\n }\n .score-pill {\n padding: 4px 5px;\n font-size: 0.68rem;\n }\n .detector-scores-wrap h3 {\n font-size: 1.15rem;\n }\n }\n<\/style>\n\n<div class=\"detector-scores-wrap\">\n <h3>Jak to detektory obrazu vyhodnotily<\/h3>\n\n <table class=\"detector-scores-table\">\n <thead>\n <tr>\n <th>Obr\u00e1zek Model<\/th>\n <th>Kategorie<\/th>\n <th>TruthScan<\/th>\n <th>Sight Engine<\/th>\n <th>Um\u011bl\u00e1 inteligence nebo ne<\/th>\n <th>Winston AI<\/th>\n <th>WasItAI<\/th>\n <\/tr>\n <\/thead>\n <tbody>\n <tr>\n <td>Nano ban\u00e1n <span style=\"font-size:0.75rem; color:#888; display:block;\">Deepfake Editace<\/span><\/td>\n <td>Podvody<\/td>\n <td><span class=\"score-pill score-pass\">99.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">98.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">99.10% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">31.98% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">99.00% AI<\/span><\/td>\n <\/tr>\n <\/tbody>\n <\/table>\n<\/div>\n\n\n\n<p style=\"font-size:17px\">Olejovit\u00fd kal, kter\u00fd Nano Banana na v\u00fdstupu vytv\u00e1\u0159ela, m\u011bl p\u0159esv\u011bd\u010div\u00fd \u0161pinav\u00fd lesk a p\u016fsobil prom\u00e1\u010den\u00fdm dojmem - v\u016fbec ne kreslen\u00fdm. Na\u0161t\u011bst\u00ed v\u011bt\u0161ina detektor\u016f, kter\u00fdmi jsem obr\u00e1zek prohnal, jej spr\u00e1vn\u011b ozna\u010dila za podvrh. Jedin\u00fdm vybo\u010duj\u00edc\u00edm detektorem byl Winston, kter\u00fd op\u011bt podal \u0161patn\u00fd v\u00fdkon a obr\u00e1zek nespr\u00e1vn\u011b klasifikoval jako autentick\u00fd. Nejp\u0159esn\u011bji tento obr\u00e1zek odhalily detektory AI or Not, TruthScan a WasItAI.<\/p>\n\n\n\n<p style=\"font-size:17px\">Je znepokojuj\u00edc\u00ed, jak jednoduch\u00e9 se zd\u00e1 nahr\u00e1t obr\u00e1zek do chatbota a nechat ho zcela zm\u011bnit. Dok\u00e1\u017eu si p\u0159edstavit, \u017ee podvodn\u00edci a p\u0159ekupn\u00edci pou\u017e\u00edvaj\u00ed n\u00e1stroje pro generov\u00e1n\u00ed obr\u00e1zk\u016f, aby se pokusili o podvodn\u00e9 reklamace vr\u00e1cen\u00ed zbo\u017e\u00ed nebo se zapojili do podvod\u016f na trhu.<\/p>\n\n\n\n<p style=\"font-size:17px\">Pokud platformy pro elektronick\u00e9 obchodov\u00e1n\u00ed vy\u017eaduj\u00ed jako d\u016fkaz pro zah\u00e1jen\u00ed reklamace po\u0161kozen\u00e9ho bal\u00edku pouze obr\u00e1zek, pak jsou bez spolehliv\u00e9ho zp\u016fsobu, jak zjistit manipulaci s um\u011blou inteligenc\u00ed, vystaveny v\u00fdznamn\u00e9mu vektoru \u00fatoku. Nap\u0159\u00edklad ofici\u00e1ln\u00ed str\u00e1nka spole\u010dnosti Amazon <a href=\"https:\/\/sellercentral.amazon.com\/help\/hub\/reference\/external\/GYW2EV9FEMJ3JJGR?locale=en-US\" target=\"_blank\" rel=\"noopener\">z\u00e1sady vracen\u00ed zbo\u017e\u00ed<\/a> doslova uv\u00e1d\u00ed<strong>, “<em>Nebezpe\u010dn\u00e9 l\u00e1tky, v\u010detn\u011b ho\u0159lav\u00fdch kapalin nebo plyn\u016f, se nevracej\u00ed.<\/em>.”<\/strong><\/p>\n\n\n\n<p style=\"font-size:17px\">Pokud byste tedy skute\u010dn\u011b obdr\u017eeli z\u00e1silku pokrytou olejem nebo mastn\u00fdm \u010dern\u00fdm kalem, nemuseli byste ji vracet, ale mohli byste dostat zp\u011bt pen\u00edze. Pokud by se v\u00e1m to z n\u011bjak\u00e9ho d\u016fvodu skute\u010dn\u011b stalo, Amazon by pravd\u011bpodobn\u011b po\u017eadoval d\u016fkaz v podob\u011b fotografie. Ch\u00e1pete, jak\u00fd probl\u00e9m mohou p\u0159edstavovat obr\u00e1zky generovan\u00e9 um\u011blou inteligenc\u00ed, kdy\u017e jsou k nerozezn\u00e1n\u00ed od skute\u010dn\u00fdch?<\/p>\n\n\n\n<h2 class=\"wp-block-heading has-text-align-left\">#4. \"Cockroach Meal\" (generov\u00e1no pomoc\u00ed Nano Banana)<\/h2>\n\n\n\n<p style=\"font-size:17px\">Dal\u0161\u00ed obr\u00e1zek um\u011bl\u00e9 inteligence v\u00e1s mo\u017en\u00e1 p\u0159iprav\u00ed o chu\u0165 k j\u00eddlu. Jak se uk\u00e1zalo, existuj\u00ed zdokumentovan\u00e9 p\u0159\u00edpady, kdy lid\u00e9 pomoc\u00ed um\u011bl\u00e9 inteligence p\u0159ij\u00edmali <a href=\"https:\/\/dailydot.com\/ai-generated-undercooked-food-doordash\" data-type=\"link\" data-id=\"https:\/\/dailydot.com\/ai-generated-undercooked-food-doordash\" target=\"_blank\" rel=\"noopener\">podvodn\u00e9 n\u00e1hrady<\/a> o aplikac\u00edch na rozvoz j\u00eddla. P\u0159i \u010dtvrt\u00e9m testu jsem postupoval stejn\u011b jako p\u0159i t\u0159et\u00edm testu. Tentokr\u00e1t jsem si vyfotil pr\u00e1zdnou krabici po ob\u011bd\u011b.<\/p>\n\n\n\n<p><strong>Zde je skute\u010dn\u00fd obr\u00e1zek:<\/strong><\/p>\n\n\n\n<figure class=\"wp-block-image size-large is-resized\"><img loading=\"lazy\" decoding=\"async\" width=\"768\" height=\"1024\" src=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/IMG_6145-768x1024.jpg\" alt=\"Fotografie pr\u00e1zdn\u00e9 krabice od j\u00eddla.\" class=\"wp-image-5296\" style=\"width:462px;height:auto\" title=\"\" srcset=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/IMG_6145-768x1024.jpg 768w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/IMG_6145-225x300.jpg 225w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/IMG_6145-scaled.jpg 1024w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/IMG_6145-9x12.jpg 9w\" sizes=\"auto, (max-width: 768px) 100vw, 768px\" \/><\/figure>\n\n\n\n<p style=\"font-size:17px\">Pot\u00e9 jsem do aplikace Nano Banana vlo\u017eil skute\u010dnou fotografii, kterou jsem po\u0159\u00eddil, a zadal j\u00ed n\u00e1sleduj\u00edc\u00ed p\u0159\u00edkaz:<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p class=\"has-vivid-cyan-blue-color has-text-color has-link-color wp-elements-e9f00ca4e893e7742bce463148a639f5\"><strong><em>\"Upravit tento obr\u00e1zek. Nic nem\u011b\u0148te, jen p\u0159idejte nudle do m\u00edsta, kde je f\u00f3lie, a p\u0159idejte do j\u00eddla hromadu mal\u00fdch \u0161v\u00e1b\u016f. A\u0165 to v\u0161echno vypad\u00e1 realisticky a ne kreslen\u011b.\" <\/em><\/strong><\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h3 class=\"wp-block-heading\">V\u00fdstup nano ban\u00e1n\u016f:<\/h3>\n\n\n\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"747\" height=\"1024\" src=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Cockroach-meal.jpg\" alt=\"Um\u011blou inteligenc\u00ed upraven\u00e1 fotografie krabice na j\u00eddlo s nudlemi a brouky uvnit\u0159 krabice\" class=\"wp-image-5295\" title=\"\" srcset=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Cockroach-meal.jpg 747w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Cockroach-meal-219x300.jpg 219w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Cockroach-meal-9x12.jpg 9w\" sizes=\"auto, (max-width: 747px) 100vw, 747px\" \/><\/figure>\n\n\n\n<style>\n @import url('https:\/\/fonts.googleapis.com\/css2?family=DM+Sans:wght@400;500;600;700&family=Outfit:wght@600;700;800&display=swap');\n\n .detector-scores-wrap {\n max-width: 780px;\n margin: 2rem auto;\n font-family: 'DM Sans', sans-serif;\n -webkit-font-smoothing: antialiased;\n }\n\n .detector-scores-wrap h3 {\n font-family: 'Outfit', sans-serif;\n font-size: 1.4rem;\n font-weight: 800;\n color: #1a1a2e;\n margin-bottom: 1.25rem;\n letter-spacing: -0.02em;\n }\n\n .detector-scores-table {\n width: 100%;\n border-collapse: separate;\n border-spacing: 0;\n border-radius: 14px;\n overflow: hidden;\n table-layout: fixed;\n box-shadow:\n 0 1px 3px rgba(0,0,0,0.06),\n 0 8px 24px rgba(0,0,0,0.08),\n 0 0 0 1px rgba(0,0,0,0.04);\n }\n\n .detector-scores-table thead th {\n background: #1a1a2e;\n color: #ffffff;\n font-family: 'Outfit', sans-serif;\n font-weight: 700;\n font-size: 0.7rem;\n text-transform: uppercase;\n letter-spacing: 0.04em;\n padding: 14px 6px;\n text-align: center;\n border: none;\n }\n\n .detector-scores-table thead th:first-child {\n text-align: left;\n padding-left: 14px;\n width: 18%;\n }\n\n .detector-scores-table thead th:nth-child(2) {\n width: 12%;\n }\n\n .detector-scores-table tbody tr {\n transition: background 0.2s ease;\n }\n\n .detector-scores-table tbody tr:hover {\n background: rgba(0,0,0,0.02);\n }\n\n .detector-scores-table tbody td {\n padding: 14px 6px;\n text-align: center;\n font-size: 0.9rem;\n font-weight: 500;\n color: #2d2d3a;\n border-bottom: 1px solid #eee;\n }\n\n .detector-scores-table tbody tr:last-child td {\n border-bottom: none;\n }\n\n .detector-scores-table tbody td:first-child {\n text-align: left;\n padding-left: 14px;\n font-weight: 600;\n color: #1a1a2e;\n }\n\n .score-pill {\n font-weight: 700;\n font-family: 'Outfit', sans-serif;\n font-size: 0.78rem;\n border-radius: 6px;\n padding: 5px 8px;\n display: inline-block;\n min-width: 0;\n text-align: center;\n white-space: nowrap;\n }\n\n .score-pass {\n background: #22c55e;\n color: #fff;\n }\n\n .score-warn {\n background: #fb923c;\n color: #fff;\n }\n\n .score-fail {\n background: #ef4444;\n color: #fff;\n }\n\n @media (max-width: 600px) {\n .detector-scores-table thead th,\n .detector-scores-table tbody td {\n padding: 10px 4px;\n font-size: 0.7rem;\n }\n .score-pill {\n padding: 4px 5px;\n font-size: 0.68rem;\n }\n .detector-scores-wrap h3 {\n font-size: 1.15rem;\n }\n }\n<\/style>\n\n<div class=\"detector-scores-wrap\">\n <h3>Jak to detektory obrazu vyhodnotily<\/h3>\n\n <table class=\"detector-scores-table\">\n <thead>\n <tr>\n <th>Obr\u00e1zek Model<\/th>\n <th>Kategorie<\/th>\n <th>TruthScan<\/th>\n <th>Sight Engine<\/th>\n <th>Um\u011bl\u00e1 inteligence nebo ne<\/th>\n <th>Winston AI<\/th>\n <th>WasItAI<\/th>\n <\/tr>\n <\/thead>\n <tbody>\n <tr>\n <td>Nano ban\u00e1n <span style=\"font-size:0.75rem; color:#888; display:block;\">Deepfake Editace<\/span><\/td>\n <td>Podvody<\/td>\n <td><span class=\"score-pill score-pass\">97.15% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">18.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-warn\">85.87% AI<\/span><\/td>\n <td><span class=\"score-pill score-warn\">72.29% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">1.00% AI<\/span><\/td>\n <\/tr>\n <\/tbody>\n <\/table>\n<\/div>\n\n\n\n<p style=\"font-size:17px\">N\u00e1stroj TruthScan op\u011bt s jistotou detekoval m\u00e9dia generovan\u00e1 um\u011blou inteligenc\u00ed. Sight Engine a WasitAI zcela selhaly. Detektor AI or Not byl p\u0159i skenov\u00e1n\u00ed tohoto obr\u00e1zku AI druh\u00fd nejp\u0159esn\u011bj\u0161\u00ed (s jistotou 85% AI) a Winston si vedl l\u00e9pe ne\u017e ve t\u0159et\u00edm testu (s jistotou 72% AI).<\/p>\n\n\n\n<p style=\"font-size:17px\">Na tomto sn\u00edmku byly v\u0161echny nudle a \u0161v\u00e1bi vytvo\u0159eny v\u00fdhradn\u011b um\u011blou inteligenc\u00ed, p\u0159esto se zd\u00e1lo, \u017ee dva z detektor\u016f si to nemysl\u00ed, a pouze jeden (TruthScan) si byl nad 90% jist\u00fd manipulac\u00ed um\u011bl\u00e9 inteligence.<\/p>\n\n\n\n<p style=\"font-size:17px\">Znepokojuj\u00edc\u00ed bylo, jak jednoduch\u00e9 bylo vytvo\u0159it tento fale\u0161n\u00fd obraz. Od nahr\u00e1n\u00ed fotografie, kterou jsem po\u0159\u00eddil, p\u0159es v\u00fdzvu Nano Banana a\u017e po z\u00edsk\u00e1n\u00ed v\u00fdstupu to trvalo jen asi 2 minuty. No\u010dn\u00ed m\u016fra pro aplikace na rozvoz j\u00eddla; sen pro hladov\u00e9 lh\u00e1\u0159e. Faktor podvodu.<\/p>\n\n\n\n<p class=\"has-vivid-red-color has-text-color has-link-color wp-elements-4c89b6ee3afcbdd6b6617c150e5b2e03\"><strong>ZVE\u0158EJN\u011aN\u00cd: N\u011bkter\u00e9 z obr\u00e1zk\u016f generovan\u00fdch um\u011blou inteligenc\u00ed, kter\u00e9 byly zahrnuty do n\u00e1sleduj\u00edc\u00edch test\u016f, m\u011bly demonstrovat, jak mohou b\u00fdt obr\u00e1zky um\u011bl\u00e9 inteligence pou\u017eity k dezinformaci. Obsahuj\u00ed kontroverzn\u00ed sn\u00edmky a citliv\u00e1 t\u00e9mata<\/strong>, <strong>a NEJSOU ur\u010deny k reprezentaci konkr\u00e9tn\u00edch politick\u00fdch<\/strong> <strong>n\u00e1zory nebo ideologie.<\/strong><\/p>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>#5<\/strong>. \"Autovandalismus\" (generov\u00e1no pomoc\u00ed Nano Banana)<\/h2>\n\n\n\n<p style=\"font-size:17px\">Na internetu jsem na\u0161el skute\u010dn\u00fd obr\u00e1zek auta na prodej. Docela v\u0161edn\u00ed. Ale co kdyby n\u011bkdo cht\u011bl tento obr\u00e1zek pou\u017e\u00edt k \u0161\u00ed\u0159en\u00ed dezinformac\u00ed nebo k pod\u00e1n\u00ed fale\u0161n\u00e9 pojistn\u00e9 ud\u00e1losti? Generativn\u00ed um\u011bl\u00e1 inteligence to bohu\u017eel umo\u017e\u0148uje snadno a rychle.<\/p>\n\n\n\n<p style=\"font-size:17px\"><strong>Pro tento p\u00e1t\u00fd test jsem vzal skute\u010dn\u00fd obr\u00e1zek auta a zadal Nano Banana n\u00e1sleduj\u00edc\u00ed v\u00fdzvu:<\/strong><\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p class=\"has-vivid-cyan-blue-color has-text-color has-link-color wp-elements-d06074fe9533ce61c08bcf9750cd398d\">“<strong><em>Na aut\u011b nic nem\u011b\u0148te, jen na n\u011bj p\u0159idejte barvu ve spreji s n\u00e1pisem \"Svobodn\u00e1 Palestina\" a praskl\u00e9 \u010deln\u00ed sklo. P\u0159edn\u00ed sv\u011btla jsou rozbit\u00e1, zrc\u00e1tko je rozbit\u00e9 a vis\u00ed dol\u016f a dv\u011b okna na boku auta jsou rozbit\u00e1.<\/em><\/strong> <strong>rozbit\u00e9<\/strong>.”<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Zde je skute\u010dn\u00fd obr\u00e1zek:<\/strong><\/h3>\n\n\n\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"940\" height=\"627\" src=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/2000_honda_civic-si_2000_honda_civic-si_bbeed9f1-7b50-4314-a504-40be9411b8ab-mXhf1y-2.webp\" alt=\"Fotografie modr\u00e9ho auta\" class=\"wp-image-5298\" title=\"\" srcset=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/2000_honda_civic-si_2000_honda_civic-si_bbeed9f1-7b50-4314-a504-40be9411b8ab-mXhf1y-2.webp 940w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/2000_honda_civic-si_2000_honda_civic-si_bbeed9f1-7b50-4314-a504-40be9411b8ab-mXhf1y-2-300x200.webp 300w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/2000_honda_civic-si_2000_honda_civic-si_bbeed9f1-7b50-4314-a504-40be9411b8ab-mXhf1y-2-768x512.webp 768w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/2000_honda_civic-si_2000_honda_civic-si_bbeed9f1-7b50-4314-a504-40be9411b8ab-mXhf1y-2-18x12.webp 18w\" sizes=\"auto, (max-width: 940px) 100vw, 940px\" \/><\/figure>\n\n\n\n<h3 class=\"wp-block-heading\">Obr\u00e1zek Nano Banana AI:<\/h3>\n\n\n\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"683\" src=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Protest-Car-Damage.jpg\" alt=\"Fotografie modr\u00e9ho vozu upraven\u00e1 um\u011blou inteligenc\u00ed, na kter\u00e9 je vid\u011bt praskl\u00e9 \u010deln\u00ed sklo a po\u0161kozen\u00ed post\u0159ikem na boku vozu.\" class=\"wp-image-5299\" title=\"\" srcset=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Protest-Car-Damage.jpg 1024w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Protest-Car-Damage-300x200.jpg 300w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Protest-Car-Damage-768x512.jpg 768w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Protest-Car-Damage-18x12.jpg 18w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<style>\n @import url('https:\/\/fonts.googleapis.com\/css2?family=DM+Sans:wght@400;500;600;700&family=Outfit:wght@600;700;800&display=swap');\n\n .detector-scores-wrap {\n max-width: 780px;\n margin: 2rem auto;\n font-family: 'DM Sans', sans-serif;\n -webkit-font-smoothing: antialiased;\n }\n\n .detector-scores-wrap h3 {\n font-family: 'Outfit', sans-serif;\n font-size: 1.4rem;\n font-weight: 800;\n color: #1a1a2e;\n margin-bottom: 1.25rem;\n letter-spacing: -0.02em;\n }\n\n .detector-scores-table {\n width: 100%;\n border-collapse: separate;\n border-spacing: 0;\n border-radius: 14px;\n overflow: hidden;\n table-layout: fixed;\n box-shadow:\n 0 1px 3px rgba(0,0,0,0.06),\n 0 8px 24px rgba(0,0,0,0.08),\n 0 0 0 1px rgba(0,0,0,0.04);\n }\n\n .detector-scores-table thead th {\n background: #1a1a2e;\n color: #ffffff;\n font-family: 'Outfit', sans-serif;\n font-weight: 700;\n font-size: 0.7rem;\n text-transform: uppercase;\n letter-spacing: 0.04em;\n padding: 14px 6px;\n text-align: center;\n border: none;\n }\n\n .detector-scores-table thead th:first-child {\n text-align: left;\n padding-left: 14px;\n width: 18%;\n }\n\n .detector-scores-table thead th:nth-child(2) {\n width: 14%;\n }\n\n .detector-scores-table tbody tr {\n transition: background 0.2s ease;\n }\n\n .detector-scores-table tbody tr:hover {\n background: rgba(0,0,0,0.02);\n }\n\n .detector-scores-table tbody td {\n padding: 14px 6px;\n text-align: center;\n font-size: 0.9rem;\n font-weight: 500;\n color: #2d2d3a;\n border-bottom: 1px solid #eee;\n }\n\n .detector-scores-table tbody tr:last-child td {\n border-bottom: none;\n }\n\n .detector-scores-table tbody td:first-child {\n text-align: left;\n padding-left: 14px;\n font-weight: 600;\n color: #1a1a2e;\n }\n\n .score-pill {\n font-weight: 700;\n font-family: 'Outfit', sans-serif;\n font-size: 0.78rem;\n border-radius: 6px;\n padding: 5px 8px;\n display: inline-block;\n min-width: 0;\n text-align: center;\n white-space: nowrap;\n }\n\n .score-pass {\n background: #22c55e;\n color: #fff;\n }\n\n .score-warn {\n background: #fb923c;\n color: #fff;\n }\n\n .score-fail {\n background: #ef4444;\n color: #fff;\n }\n\n @media (max-width: 600px) {\n .detector-scores-table thead th,\n .detector-scores-table tbody td {\n padding: 10px 4px;\n font-size: 0.7rem;\n }\n .score-pill {\n padding: 4px 5px;\n font-size: 0.68rem;\n }\n .detector-scores-wrap h3 {\n font-size: 1.15rem;\n }\n }\n<\/style>\n\n<div class=\"detector-scores-wrap\">\n <h3>Jak to detektory obrazu vyhodnotily<\/h3>\n\n <table class=\"detector-scores-table\">\n <thead>\n <tr>\n <th>Obr\u00e1zek Model<\/th>\n <th>Kategorie<\/th>\n <th>TruthScan<\/th>\n <th>Sight Engine<\/th>\n <th>Um\u011bl\u00e1 inteligence nebo ne<\/th>\n <th>Winston AI<\/th>\n <th>WasItAI<\/th>\n <\/tr>\n <\/thead>\n <tbody>\n <tr>\n <td>Nano ban\u00e1n <span style=\"font-size:0.75rem; color:#888; display:block;\">Deepfake Editace<\/span><\/td>\n <td style=\"font-size:0.82rem;\">Podvody \/ dezinformace<\/td>\n <td><span class=\"score-pill score-pass\">97.48% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">18.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-warn\">89.52% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">0.02% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">99.00% AI<\/span><\/td>\n <\/tr>\n <\/tbody>\n <\/table>\n<\/div>\n\n\n\n<p style=\"font-size:17px\">V\u00fdsledky detektor\u016f pro tento v\u00fdstup byly sm\u00ed\u0161en\u00e9. V tomto testu dos\u00e1hl WasItAI (p\u0159ekvapiv\u011b) nejvy\u0161\u0161\u00edho sk\u00f3re detekce AI (99%). TruthScan m\u011bl druh\u00e9 nejvy\u0161\u0161\u00ed sk\u00f3re (97%). Sight Engine v\u00fdrazn\u011b selhal p\u0159i detekci syntetick\u00fdch prvk\u016f v obr\u00e1zku a Winston si vedl nejh\u016f\u0159e ze v\u0161ech detektor\u016f (p\u0159i\u0159adil \u0161anci na zapojen\u00ed men\u0161\u00ed ne\u017e 1%). AI or Not si vedl dob\u0159e, ale p\u0159esnost byla st\u00e1le ni\u017e\u0161\u00ed ne\u017e 90%.<br><br>Promluvme si o obr\u00e1zku. Dva nejznepokojiv\u011bj\u0161\u00ed aspekty takov\u00e9ho obsahu vytvo\u0159en\u00e9ho um\u011blou inteligenc\u00ed nebo pozm\u011bn\u011bn\u00e9ho obsahu spo\u010d\u00edvaj\u00ed v tom, \u017ee oba zahrnuj\u00ed klam\u00e1n\u00ed. Za prv\u00e9, pou\u017eit\u00ed um\u011bl\u00e9 inteligence k \u00faprav\u011b b\u011b\u017en\u00fdch obr\u00e1zk\u016f za \u00fa\u010delem vytvo\u0159en\u00ed politicky podbarven\u00e9ho obsahu je rychl\u00e9, snadn\u00e9 a vypad\u00e1 re\u00e1ln\u011b. A\u0165 u\u017e ti, kdo tento typ obsahu vytv\u00e1\u0159ej\u00ed, usiluj\u00ed pouze o anga\u017eovanost, nebo jsou politick\u00fdmi provokat\u00e9ry, probl\u00e9m je v tom, \u017ee jde o fale\u0161n\u00e9 informace, kter\u00e9 jsou prezentov\u00e1ny jako autentick\u00e9.<\/p>\n\n\n\n<p style=\"font-size:17px\">Druh\u00fdm zp\u016fsobem, jak lze podobn\u00e9 n\u00e1stroje pro generov\u00e1n\u00ed obr\u00e1zk\u016f zneu\u017e\u00edt, je podvod. Pokud chce podvodn\u00edk podat fale\u0161nou \u017e\u00e1dost o poji\u0161t\u011bn\u00ed auta, mus\u00ed p\u0159edlo\u017eit fale\u0161n\u00e9 d\u016fkazy. M\u00edsto toho, aby tr\u00e1vili hodiny ve Photoshopu \u00fapravami obr\u00e1zk\u016f a upravov\u00e1n\u00edm d\u016fkaz\u016f, mohou pou\u017e\u00edt gener\u00e1tory obr\u00e1zk\u016f k rychl\u00e9 v\u00fdrob\u011b fale\u0161n\u00fdch d\u016fkaz\u016f.<\/p>\n\n\n\n<p style=\"font-size:17px\">V p\u0159\u00edkladu s um\u011blou inteligenc\u00ed v\u00fd\u0161e si v\u0161imn\u011bte, jak by podvodn\u00edk mohl pou\u017e\u00edt podobn\u00fd postup k pod\u00e1n\u00ed fale\u0161n\u00e9 \u017e\u00e1dosti podle spole\u010dnosti Progressive. <a href=\"https:\/\/www.progressive.com\/answers\/does-car-insurance-cover-vandalism\/\" data-type=\"link\" data-id=\"https:\/\/www.progressive.com\/answers\/does-car-insurance-cover-vandalism\/\" target=\"_blank\" rel=\"noopener\">Politika komplexn\u00edho kryt\u00ed<\/a>, kter\u00e1 se konkr\u00e9tn\u011b vztahuje na:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Roz\u0159ezan\u00e9 nebo <a href=\"https:\/\/www.progressive.com\/answers\/does-car-insurance-cover-tire-damage\/\" target=\"_blank\" rel=\"noopener\">po\u0161kozen\u00e9 pneumatiky<\/a><\/strong><\/li>\n\n\n\n<li><strong>Rozbit\u00e1 okna, p\u0159edn\u00ed nebo zadn\u00ed sv\u011btla<\/strong><\/li>\n\n\n\n<li><strong>Po\u0161kozen\u00ed laku ve spreji<\/strong><\/li>\n\n\n\n<li><strong><a href=\"https:\/\/www.progressive.com\/answers\/does-car-insurance-cover-scratches\/\" target=\"_blank\" rel=\"noopener\">Prom\u00e1\u010dkliny nebo \u0161kr\u00e1bance<\/a> od n\u011bkoho, kdo v\u00e1m odemkne auto kl\u00ed\u010dem.<\/strong><\/li>\n\n\n\n<li><strong>P\u0159id\u00e1v\u00e1n\u00ed cukru nebo jin\u00fdch l\u00e1tek do n\u00e1dr\u017ee na benzin<\/strong><\/li>\n<\/ul>\n\n\n\n<p>P\u0159edkl\u00e1d\u00e1n\u00ed fale\u0161n\u00fdch d\u016fkaz\u016f je samoz\u0159ejm\u011b nejen klamav\u00e9, ale tak\u00e9 zcela nez\u00e1konn\u00e9. V ide\u00e1ln\u00edm sv\u011bt\u011b by nikdo neporu\u0161oval z\u00e1kon, ve skute\u010dnosti se v\u0161ak trestn\u00e9 \u010diny d\u011bj\u00ed ka\u017ed\u00fd den. A bez spolehliv\u00e9ho odhalen\u00ed hroz\u00ed, \u017ee automobilov\u00e9 poji\u0161\u0165ovny budou o\u0161izeny o miliony, proto\u017ee podvodn\u00fdch pojistn\u00fdch ud\u00e1lost\u00ed s vyu\u017eit\u00edm um\u011bl\u00e9 inteligence p\u0159ib\u00fdv\u00e1.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">#6. \"Soda Musk a Don\" (generov\u00e1no pomoc\u00ed Nano Banana)<\/h2>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"559\" src=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Elon-Spilling-Soda-1024x559.jpeg\" alt=\"Obr\u00e1zek Elona Muska a Donalda Trumpa vytvo\u0159en\u00fd um\u011blou inteligenc\u00ed\" class=\"wp-image-5300\" title=\"\" srcset=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Elon-Spilling-Soda-1024x559.jpeg 1024w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Elon-Spilling-Soda-300x164.jpeg 300w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Elon-Spilling-Soda-768x419.jpeg 768w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Elon-Spilling-Soda-1536x838.jpeg 1536w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Elon-Spilling-Soda-2048x1117.jpeg 1877w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Elon-Spilling-Soda-18x10.jpeg 18w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3-Elon-Spilling-Soda-scaled.jpeg 1878w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<style>\n @import url('https:\/\/fonts.googleapis.com\/css2?family=DM+Sans:wght@400;500;600;700&family=Outfit:wght@600;700;800&display=swap');\n\n .detector-scores-wrap {\n max-width: 780px;\n margin: 2rem auto;\n font-family: 'DM Sans', sans-serif;\n -webkit-font-smoothing: antialiased;\n }\n\n .detector-scores-wrap h3 {\n font-family: 'Outfit', sans-serif;\n font-size: 1.4rem;\n font-weight: 800;\n color: #1a1a2e;\n margin-bottom: 1.25rem;\n letter-spacing: -0.02em;\n }\n\n .detector-scores-table {\n width: 100%;\n border-collapse: separate;\n border-spacing: 0;\n border-radius: 14px;\n overflow: hidden;\n table-layout: fixed;\n box-shadow:\n 0 1px 3px rgba(0,0,0,0.06),\n 0 8px 24px rgba(0,0,0,0.08),\n 0 0 0 1px rgba(0,0,0,0.04);\n }\n\n .detector-scores-table thead th {\n background: #1a1a2e;\n color: #ffffff;\n font-family: 'Outfit', sans-serif;\n font-weight: 700;\n font-size: 0.7rem;\n text-transform: uppercase;\n letter-spacing: 0.04em;\n padding: 14px 6px;\n text-align: center;\n border: none;\n }\n\n .detector-scores-table thead th:first-child {\n text-align: left;\n padding-left: 14px;\n width: 18%;\n }\n\n .detector-scores-table thead th:nth-child(2) {\n width: 14%;\n }\n\n .detector-scores-table tbody tr {\n transition: background 0.2s ease;\n }\n\n .detector-scores-table tbody tr:hover {\n background: rgba(0,0,0,0.02);\n }\n\n .detector-scores-table tbody td {\n padding: 14px 6px;\n text-align: center;\n font-size: 0.9rem;\n font-weight: 500;\n color: #2d2d3a;\n border-bottom: 1px solid #eee;\n }\n\n .detector-scores-table tbody tr:last-child td {\n border-bottom: none;\n }\n\n .detector-scores-table tbody td:first-child {\n text-align: left;\n padding-left: 14px;\n font-weight: 600;\n color: #1a1a2e;\n }\n\n .score-pill {\n font-weight: 700;\n font-family: 'Outfit', sans-serif;\n font-size: 0.78rem;\n border-radius: 6px;\n padding: 5px 8px;\n display: inline-block;\n min-width: 0;\n text-align: center;\n white-space: nowrap;\n }\n\n .score-pass {\n background: #22c55e;\n color: #fff;\n }\n\n .score-warn {\n background: #fb923c;\n color: #fff;\n }\n\n .score-fail {\n background: #ef4444;\n color: #fff;\n }\n\n @media (max-width: 600px) {\n .detector-scores-table thead th,\n .detector-scores-table tbody td {\n padding: 10px 4px;\n font-size: 0.7rem;\n }\n .score-pill {\n padding: 4px 5px;\n font-size: 0.68rem;\n }\n .detector-scores-wrap h3 {\n font-size: 1.15rem;\n }\n }\n<\/style>\n\n<div class=\"detector-scores-wrap\">\n <h3>Jak to detektory obrazu vyhodnotily<\/h3>\n\n <table class=\"detector-scores-table\">\n <thead>\n <tr>\n <th>Obr\u00e1zek Model<\/th>\n <th>Kategorie<\/th>\n <th>TruthScan<\/th>\n <th>Sight Engine<\/th>\n <th>Um\u011bl\u00e1 inteligence nebo ne<\/th>\n <th>Winston AI<\/th>\n <th>WasItAI<\/th>\n <\/tr>\n <\/thead>\n <tbody>\n <tr>\n <td>Nano ban\u00e1n <span style=\"font-size:0.75rem; color:#888; display:block;\">Pln\u00e1 generace<\/span><\/td>\n <td style=\"font-size:0.82rem;\">Disinfo<\/td>\n <td><span class=\"score-pill score-pass\">99.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">99.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">99.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">0.24% AI<\/span><\/td>\n <td><span class=\"score-pill score-warn\">86.00% AI<\/span><\/td>\n <\/tr>\n <\/tbody>\n <\/table>\n<\/div>\n\n\n\n<p style=\"font-size:17px\">Pro tento test jsem op\u011bt pou\u017eil Nano Banana a c\u00edlem bylo p\u0159edv\u00e9st fale\u0161n\u00e9, politicky nabit\u00e9 sn\u00edmky s hloup\u011bj\u0161\u00edm podtextem. Programy TruthScan, Sight Engine a AIorNot ozna\u010dily vygenerovan\u00fd obr\u00e1zek s hodnocen\u00edm 99% detekce AI. WasitAI si nebyl tak jist\u00fd - detekoval obr\u00e1zek vygenerovan\u00fd AI s jistotou pouze 86%. A Winston zcela selhal a p\u0159i\u0159adil AI sk\u00f3re 0,2%. Pro vygenerov\u00e1n\u00ed obr\u00e1zku jsem zadal Nano Bananovi tuto v\u00fdzvu:<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p class=\"has-vivid-cyan-blue-color has-text-color has-link-color wp-elements-4f7e002ecaefec24bbebdca3101fff42\"><strong>“<em>Vygenerujte fotku Elona Muska, jak si vyl\u00e9v\u00e1 velk\u00fd kel\u00edmek limon\u00e1dy jako Big Gulp na ko\u0161ili a vy\u0161iluje, zat\u00edmco vedle n\u011bj sed\u00ed rozesm\u00e1t\u00fd Donald Trump. Jsou v tryskov\u00e9m letadle. To je ta sc\u00e9na. A fotka by m\u011bla vypadat realisticky, jako by byla po\u0159\u00edzena fotoapar\u00e1tem na momentky, a m\u011bla by obsahovat artefakty, jako je blesk fotoapar\u00e1tu odr\u00e1\u017eej\u00edc\u00ed se od okna letadla a mimo letadlo. Je noc. M\u011bla by vypadat hyperrealisticky.<\/em>“<\/strong><\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p style=\"font-size:17px\">Kdykoli n\u011bkter\u00fd z t\u011bchto program\u016f nedok\u00e1\u017ee rozpoznat um\u011blou inteligenci, ozna\u010d\u00ed ji za pravou. Jakkoli je tento p\u0159\u00edklad obr\u00e1zku z testu hloup\u00fd, nen\u00ed nic vtipn\u00e9ho na tom, kdy\u017e si mysl\u00edte, \u017ee deepfake je prav\u00fd.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">#7. \"Still Alive\" (generuje Nano Banana)<\/h2>\n\n\n\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"765\" height=\"1024\" src=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3Christian_Ransom_Photo.jpg\" alt=\"Obr\u00e1zek mu\u017ee s n\u00e1pisem "help" vytvo\u0159en\u00fd um\u011blou inteligenc\u00ed.\" class=\"wp-image-5301\" title=\"\" srcset=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3Christian_Ransom_Photo.jpg 765w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3Christian_Ransom_Photo-224x300.jpg 224w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/GV3Christian_Ransom_Photo-9x12.jpg 9w\" sizes=\"auto, (max-width: 765px) 100vw, 765px\" \/><\/figure>\n\n\n\n<style>\n @import url('https:\/\/fonts.googleapis.com\/css2?family=DM+Sans:wght@400;500;600;700&family=Outfit:wght@600;700;800&display=swap');\n\n .detector-scores-wrap {\n max-width: 780px;\n margin: 2rem auto;\n font-family: 'DM Sans', sans-serif;\n -webkit-font-smoothing: antialiased;\n }\n\n .detector-scores-wrap h3 {\n font-family: 'Outfit', sans-serif;\n font-size: 1.4rem;\n font-weight: 800;\n color: #1a1a2e;\n margin-bottom: 1.25rem;\n letter-spacing: -0.02em;\n }\n\n .detector-scores-table {\n width: 100%;\n border-collapse: separate;\n border-spacing: 0;\n border-radius: 14px;\n overflow: hidden;\n table-layout: fixed;\n box-shadow:\n 0 1px 3px rgba(0,0,0,0.06),\n 0 8px 24px rgba(0,0,0,0.08),\n 0 0 0 1px rgba(0,0,0,0.04);\n }\n\n .detector-scores-table thead th {\n background: #1a1a2e;\n color: #ffffff;\n font-family: 'Outfit', sans-serif;\n font-weight: 700;\n font-size: 0.7rem;\n text-transform: uppercase;\n letter-spacing: 0.04em;\n padding: 14px 6px;\n text-align: center;\n border: none;\n }\n\n .detector-scores-table thead th:first-child {\n text-align: left;\n padding-left: 14px;\n width: 18%;\n }\n\n .detector-scores-table thead th:nth-child(2) {\n width: 14%;\n }\n\n .detector-scores-table tbody tr {\n transition: background 0.2s ease;\n }\n\n .detector-scores-table tbody tr:hover {\n background: rgba(0,0,0,0.02);\n }\n\n .detector-scores-table tbody td {\n padding: 14px 6px;\n text-align: center;\n font-size: 0.9rem;\n font-weight: 500;\n color: #2d2d3a;\n border-bottom: 1px solid #eee;\n }\n\n .detector-scores-table tbody tr:last-child td {\n border-bottom: none;\n }\n\n .detector-scores-table tbody td:first-child {\n text-align: left;\n padding-left: 14px;\n font-weight: 600;\n color: #1a1a2e;\n }\n\n .score-pill {\n font-weight: 700;\n font-family: 'Outfit', sans-serif;\n font-size: 0.78rem;\n border-radius: 6px;\n padding: 5px 8px;\n display: inline-block;\n min-width: 0;\n text-align: center;\n white-space: nowrap;\n }\n\n .score-pass {\n background: #22c55e;\n color: #fff;\n }\n\n .score-warn {\n background: #fb923c;\n color: #fff;\n }\n\n .score-fail {\n background: #ef4444;\n color: #fff;\n }\n\n @media (max-width: 600px) {\n .detector-scores-table thead th,\n .detector-scores-table tbody td {\n padding: 10px 4px;\n font-size: 0.7rem;\n }\n .score-pill {\n padding: 4px 5px;\n font-size: 0.68rem;\n }\n .detector-scores-wrap h3 {\n font-size: 1.15rem;\n }\n }\n<\/style>\n\n<div class=\"detector-scores-wrap\">\n <h3>Jak to detektory obrazu vyhodnotily<\/h3>\n\n <table class=\"detector-scores-table\">\n <thead>\n <tr>\n <th>Obr\u00e1zek Model<\/th>\n <th>Kategorie<\/th>\n <th>TruthScan<\/th>\n <th>Sight Engine<\/th>\n <th>Um\u011bl\u00e1 inteligence nebo ne<\/th>\n <th>Winston AI<\/th>\n <th>WasItAI<\/th>\n <\/tr>\n <\/thead>\n <tbody>\n <tr>\n <td>Nano ban\u00e1n <span style=\"font-size:0.75rem; color:#888; display:block;\">Pln\u00e1 generace<\/span><\/td>\n <td style=\"font-size:0.82rem;\">Deepfake<\/td>\n <td><span class=\"score-pill score-pass\">97.49% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">95.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">99.27% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">0.15% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">1.00% AI<\/span><\/td>\n <\/tr>\n <\/tbody>\n <\/table>\n<\/div>\n\n\n\n<p style=\"font-size:17px\">P\u0159ipad\u00e1 v\u00e1m tento obr\u00e1zek znepokojiv\u00fd? K vytvo\u0159en\u00ed tohoto deepfake jsem pou\u017eil Nano Banana a dal mu sv\u016fj obr\u00e1zek. Pot\u00e9 jsem mu zadal dlouhou v\u00fdzvu, kter\u00e1 m\u011b m\u011bla dostat do situace, kter\u00e1 nazna\u010duje, \u017ee jsem byl unesen. Mo\u017en\u00e1 je to proto, \u017ee tato fotografie ukazuje <em>j\u00e1<\/em> v kompromituj\u00edc\u00ed pozici. Co kdyby n\u011bkdo n\u011bco takov\u00e9ho vyrobil, poslal to m\u00e9 matce a vyhro\u017eoval, \u017ee mi ubl\u00ed\u017e\u00ed, pokud mu nepo\u0161le pen\u00edze? D\u011bsiv\u00e9 v\u011bci. <\/p>\n\n\n\n<p class=\"has-small-font-size\"><em>(\u00daplnou v\u00fdzvu, kterou jsem k tomu pou\u017eil, najdete v \u00fapln\u00e9m souboru testovac\u00edch dat na konci tohoto \u010dl\u00e1nku.)<\/em><\/p>\n\n\n\n<p style=\"font-size:17px\"><br>Na\u0161t\u011bst\u00ed pro tento sn\u00edmek, p\u0159esto\u017ee vypad\u00e1 velmi realisticky, <strong>t\u0159i z p\u011bti detektor\u016f je identifikovaly jako vytvo\u0159en\u00e9 um\u011blou inteligenc\u00ed.<\/strong> <\/p>\n\n\n\n<p style=\"font-size:17px\">AI or Not dos\u00e1hla siln\u00e9ho sk\u00f3re 99%, TruthScan 97% a Sight Engine 95%. Bohu\u017eel Winston a WasItAI klasifikovaly obr\u00e1zek jako skute\u010dn\u00fd.<\/p>\n\n\n\n<p style=\"font-size:17px\">M\u011bl bych up\u0159esnit, \u017ee tyto testy nebyly provedeny na m\u00e9m osobn\u00edm \u00fa\u010dtu Gemini. Nano Banana m\u011b nikdy nepo\u017e\u00e1dala, abych prok\u00e1zal, \u017ee jsem osoba na obr\u00e1zku. Kdokoli si mohl st\u00e1hnout moji fotografii z internetu, vlo\u017eit ji do programu Nano Banana a nechat jej vytvo\u0159it tento typ obr\u00e1zku. <\/p>\n\n\n\n<h2 class=\"wp-block-heading\">#8. \"Abuse of a Ben\" (generov\u00e1no pomoc\u00ed Nano Banana)<\/h2>\n\n\n<div class=\"wp-block-image\">\n<figure class=\"aligncenter size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"200\" height=\"200\" src=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/c945d528-730f-48c1-bdca-30e5799fc1aa_thumb-1.jpg\" alt=\"\" class=\"wp-image-5307\" title=\"\" srcset=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/c945d528-730f-48c1-bdca-30e5799fc1aa_thumb-1.jpg 200w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/c945d528-730f-48c1-bdca-30e5799fc1aa_thumb-1-150x150.jpg 150w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/c945d528-730f-48c1-bdca-30e5799fc1aa_thumb-1-12x12.jpg 12w\" sizes=\"auto, (max-width: 200px) 100vw, 200px\" \/><\/figure>\n<\/div>\n\n\n<p style=\"font-size:17px\">Seznamte se s Benem. V\u011bt\u0161inu \u010dasu tr\u00e1v\u00ed \u0159\u00edzen\u00edm r\u016fzn\u00fdch odd\u011blen\u00ed ve spole\u010dnostech, ve kter\u00fdch pracuje. Na obr\u00e1zku v\u00fd\u0161e je Benova (skute\u010dn\u00e1) \u0161vih\u00e1ck\u00e1 firemn\u00ed fotografie. Pokud to nepozn\u00e1te, Ben je profesion\u00e1l. A\u0165 u\u017e m\u00e1 jak\u00e9koli politick\u00e9 n\u00e1zory, nech\u00e1v\u00e1 si je pro sebe a jeho pracovn\u00ed rozhovory se t\u00fdkaj\u00ed jen pr\u00e1ce. Kdykoli v\u0161ak m\u016f\u017ee b\u00fdt Benova podobizna pou\u017eita k \u0161\u00ed\u0159en\u00ed politick\u00e9 propagandy nebo zpr\u00e1v bez jeho souhlasu.<\/p>\n\n\n\n<p style=\"font-size:17px\">V posledn\u00ed dob\u011b se hodn\u011b p\u00ed\u0161e o tom, jak se na internetu \u0161\u00ed\u0159\u00ed videa generovan\u00e1 um\u011blou inteligenc\u00ed, kter\u00e1 zahrnuj\u00ed agenty ICE a protesty. Na z\u00e1klad\u011b v\u0161ech dosavadn\u00edch test\u016f jsem dosp\u011bl k z\u00e1v\u011bru, \u017ee vytvo\u0159it fale\u0161nou politickou propagandu pomoc\u00ed AI je opravdu snadn\u00e9. Rozhodl jsem se testovat znovu, tentokr\u00e1t s vyu\u017eit\u00edm Bena. <\/p>\n\n\n\n<p style=\"font-size:17px\">S Benov\u00fdm svolen\u00edm jsem pou\u017eil jeho fotku hlavy k vytvo\u0159en\u00ed deepfake. <strong>Zde je obr\u00e1zek vytvo\u0159en\u00fd um\u011blou inteligenc\u00ed, kter\u00fd vytvo\u0159ila spole\u010dnost Nano Banana:<\/strong><\/p>\n\n\n<div class=\"wp-block-image\">\n<figure class=\"aligncenter size-full is-resized\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"1024\" src=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/Gemini_Generated_Image_qd4x5sqd4x5sqd4x.jpg\" alt=\"Um\u011blou inteligenc\u00ed upraven\u00e1 fotografie mu\u017ee jm\u00e9nem Ben, kter\u00fd dr\u017e\u00ed ceduli s n\u00e1pisem Abolish ICE.\" class=\"wp-image-5308\" style=\"width:537px;height:auto\" title=\"\" srcset=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/Gemini_Generated_Image_qd4x5sqd4x5sqd4x.jpg 1024w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/Gemini_Generated_Image_qd4x5sqd4x5sqd4x-300x300.jpg 300w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/Gemini_Generated_Image_qd4x5sqd4x5sqd4x-150x150.jpg 150w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/Gemini_Generated_Image_qd4x5sqd4x5sqd4x-768x768.jpg 768w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/Gemini_Generated_Image_qd4x5sqd4x5sqd4x-12x12.jpg 12w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n<\/div>\n\n\n<style>\n @import url('https:\/\/fonts.googleapis.com\/css2?family=DM+Sans:wght@400;500;600;700&family=Outfit:wght@600;700;800&display=swap');\n\n .detector-scores-wrap {\n max-width: 780px;\n margin: 2rem auto;\n font-family: 'DM Sans', sans-serif;\n -webkit-font-smoothing: antialiased;\n }\n\n .detector-scores-wrap h3 {\n font-family: 'Outfit', sans-serif;\n font-size: 1.4rem;\n font-weight: 800;\n color: #1a1a2e;\n margin-bottom: 1.25rem;\n letter-spacing: -0.02em;\n }\n\n .detector-scores-table {\n width: 100%;\n border-collapse: separate;\n border-spacing: 0;\n border-radius: 14px;\n overflow: hidden;\n table-layout: fixed;\n box-shadow:\n 0 1px 3px rgba(0,0,0,0.06),\n 0 8px 24px rgba(0,0,0,0.08),\n 0 0 0 1px rgba(0,0,0,0.04);\n }\n\n .detector-scores-table thead th {\n background: #1a1a2e;\n color: #ffffff;\n font-family: 'Outfit', sans-serif;\n font-weight: 700;\n font-size: 0.7rem;\n text-transform: uppercase;\n letter-spacing: 0.04em;\n padding: 14px 6px;\n text-align: center;\n border: none;\n }\n\n .detector-scores-table thead th:first-child {\n text-align: left;\n padding-left: 14px;\n width: 18%;\n }\n\n .detector-scores-table thead th:nth-child(2) {\n width: 14%;\n }\n\n .detector-scores-table tbody tr {\n transition: background 0.2s ease;\n }\n\n .detector-scores-table tbody tr:hover {\n background: rgba(0,0,0,0.02);\n }\n\n .detector-scores-table tbody td {\n padding: 14px 6px;\n text-align: center;\n font-size: 0.9rem;\n font-weight: 500;\n color: #2d2d3a;\n border-bottom: 1px solid #eee;\n }\n\n .detector-scores-table tbody tr:last-child td {\n border-bottom: none;\n }\n\n .detector-scores-table tbody td:first-child {\n text-align: left;\n padding-left: 14px;\n font-weight: 600;\n color: #1a1a2e;\n }\n\n .score-pill {\n font-weight: 700;\n font-family: 'Outfit', sans-serif;\n font-size: 0.78rem;\n border-radius: 6px;\n padding: 5px 8px;\n display: inline-block;\n min-width: 0;\n text-align: center;\n white-space: nowrap;\n }\n\n .score-pass {\n background: #22c55e;\n color: #fff;\n }\n\n .score-warn {\n background: #fb923c;\n color: #fff;\n }\n\n .score-fail {\n background: #ef4444;\n color: #fff;\n }\n\n @media (max-width: 600px) {\n .detector-scores-table thead th,\n .detector-scores-table tbody td {\n padding: 10px 4px;\n font-size: 0.7rem;\n }\n .score-pill {\n padding: 4px 5px;\n font-size: 0.68rem;\n }\n .detector-scores-wrap h3 {\n font-size: 1.15rem;\n }\n }\n<\/style>\n\n<div class=\"detector-scores-wrap\">\n <h3>Jak to detektory obrazu vyhodnotily<\/h3>\n\n <table class=\"detector-scores-table\">\n <thead>\n <tr>\n <th>Obr\u00e1zek Model<\/th>\n <th>Kategorie<\/th>\n <th>TruthScan<\/th>\n <th>Sight Engine<\/th>\n <th>Um\u011bl\u00e1 inteligence nebo ne<\/th>\n <th>Winston AI<\/th>\n <th>WasItAI<\/th>\n <\/tr>\n <\/thead>\n <tbody>\n <tr>\n <td>Nano ban\u00e1n<\/td>\n <td style=\"font-size:0.78rem;\">Deepfake \/ Disinfo<\/td>\n <td><span class=\"score-pill score-pass\">99.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">97.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">98.88% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">5.74% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">99.00% AI<\/span><\/td>\n <\/tr>\n <\/tbody>\n <\/table>\n<\/div>\n\n\n\n<p style=\"font-size:17px\">Trvalo to t\u0159icet sekund. Sta\u010dilo p\u0159et\u00e1hnout jeho fotku hlavy do Gemini a vyzvat ho, aby \"zvedl n\u00e1pis \"Zru\u0161te ICE\"\". V\u00fdstup z Nano Banana nen\u00ed tak fotorealistick\u00fd jako ostatn\u00ed, ale st\u00e1le vypad\u00e1 dostate\u010dn\u011b autenticky, aby byl p\u0159esv\u011bd\u010div\u00fd.<\/p>\n\n\n\n<p style=\"font-size:17px\"><span style=\"box-sizing: border-box; margin: 0px; padding: 0px;\">Na\u0161t\u011bst\u00ed \u010dty\u0159i z detektor\u016f p\u0159esn\u011b identifikovaly obraz:\u00a0<strong>TruthScan<\/strong> (9<\/span>9% AI), Was It AI (99% AI), Sight Engine (97% AI) a AI or Not (97% AI). Winston AI neusp\u011bl a p\u0159id\u011blil pouze sk\u00f3re 5% AI.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">#9. \"Club Void\" (generuje Midjourney)<\/h2>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"771\" src=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/MJ-Girl-in-club-1024x771.jpg\" alt=\"Usm\u011bvav\u00e1 \u017eena v b\u00edl\u00fdch \u0161atech s kr\u00e1tk\u00fdmi ruk\u00e1vy stoj\u00ed v tlumen\u011b osv\u011btlen\u00e9, p\u0159epln\u011bn\u00e9 m\u00edstnosti. Fotografie se zd\u00e1 b\u00fdt po\u0159\u00edzena s bleskem, co\u017e vytv\u00e1\u0159\u00ed mlhavou, namodralou atmosf\u00e9ru se slab\u00fdmi \u010d\u00e1ste\u010dkami viditeln\u00fdmi ve vzduchu. V pozad\u00ed pod po\u0161kozen\u00fdm nebo pr\u016fmyslov\u011b vypadaj\u00edc\u00edm stropem stoj\u00ed dal\u0161\u00ed lid\u00e9 v b\u011b\u017en\u00e9m oble\u010den\u00ed.\" class=\"wp-image-5309\" title=\"\" srcset=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/MJ-Girl-in-club-1024x771.jpg 1024w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/MJ-Girl-in-club-300x226.jpg 300w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/MJ-Girl-in-club-768x578.jpg 768w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/MJ-Girl-in-club-16x12.jpg 16w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/MJ-Girl-in-club.jpg 1232w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<style>\n @import url('https:\/\/fonts.googleapis.com\/css2?family=DM+Sans:wght@400;500;600;700&family=Outfit:wght@600;700;800&display=swap');\n\n .detector-scores-wrap {\n max-width: 780px;\n margin: 2rem auto;\n font-family: 'DM Sans', sans-serif;\n -webkit-font-smoothing: antialiased;\n }\n\n .detector-scores-wrap h3 {\n font-family: 'Outfit', sans-serif;\n font-size: 1.4rem;\n font-weight: 800;\n color: #1a1a2e;\n margin-bottom: 1.25rem;\n letter-spacing: -0.02em;\n }\n\n .detector-scores-table {\n width: 100%;\n border-collapse: separate;\n border-spacing: 0;\n border-radius: 14px;\n overflow: hidden;\n table-layout: fixed;\n box-shadow:\n 0 1px 3px rgba(0,0,0,0.06),\n 0 8px 24px rgba(0,0,0,0.08),\n 0 0 0 1px rgba(0,0,0,0.04);\n }\n\n .detector-scores-table thead th {\n background: #1a1a2e;\n color: #ffffff;\n font-family: 'Outfit', sans-serif;\n font-weight: 700;\n font-size: 0.7rem;\n text-transform: uppercase;\n letter-spacing: 0.04em;\n padding: 14px 6px;\n text-align: center;\n border: none;\n }\n\n .detector-scores-table thead th:first-child {\n text-align: left;\n padding-left: 14px;\n width: 18%;\n }\n\n .detector-scores-table thead th:nth-child(2) {\n width: 14%;\n }\n\n .detector-scores-table tbody tr {\n transition: background 0.2s ease;\n }\n\n .detector-scores-table tbody tr:hover {\n background: rgba(0,0,0,0.02);\n }\n\n .detector-scores-table tbody td {\n padding: 14px 6px;\n text-align: center;\n font-size: 0.9rem;\n font-weight: 500;\n color: #2d2d3a;\n border-bottom: 1px solid #eee;\n }\n\n .detector-scores-table tbody tr:last-child td {\n border-bottom: none;\n }\n\n .detector-scores-table tbody td:first-child {\n text-align: left;\n padding-left: 14px;\n font-weight: 600;\n color: #1a1a2e;\n }\n\n .score-pill {\n font-weight: 700;\n font-family: 'Outfit', sans-serif;\n font-size: 0.78rem;\n border-radius: 6px;\n padding: 5px 8px;\n display: inline-block;\n min-width: 0;\n text-align: center;\n white-space: nowrap;\n }\n\n .score-pass {\n background: #22c55e;\n color: #fff;\n }\n\n .score-warn {\n background: #fb923c;\n color: #fff;\n }\n\n .score-fail {\n background: #ef4444;\n color: #fff;\n }\n\n @media (max-width: 600px) {\n .detector-scores-table thead th,\n .detector-scores-table tbody td {\n padding: 10px 4px;\n font-size: 0.7rem;\n }\n .score-pill {\n padding: 4px 5px;\n font-size: 0.68rem;\n }\n .detector-scores-wrap h3 {\n font-size: 1.15rem;\n }\n }\n<\/style>\n\n<div class=\"detector-scores-wrap\">\n <h3>Jak to detektory obrazu vyhodnotily<\/h3>\n\n <table class=\"detector-scores-table\">\n <thead>\n <tr>\n <th>Obr\u00e1zek Model<\/th>\n <th>Kategorie<\/th>\n <th>TruthScan<\/th>\n <th>Sight Engine<\/th>\n <th>Um\u011bl\u00e1 inteligence nebo ne<\/th>\n <th>Winston AI<\/th>\n <th>WasItAI<\/th>\n <\/tr>\n <\/thead>\n <tbody>\n <tr>\n <td>MidJourney <span style=\"font-size:0.75rem; color:#888; display:block;\">Pln\u00e1 generace<\/span><\/td>\n <td>Obecn\u00e9<\/td>\n <td><span class=\"score-pill score-pass\">99.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">99.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-warn\">89.21% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">1.56% AI<\/span><\/td>\n <td><span class=\"score-pill score-warn\">70.00% AI<\/span><\/td>\n <\/tr>\n <\/tbody>\n <\/table>\n<\/div>\n\n\n\n<p style=\"font-size:17px\">V\u00fd\u0161e uveden\u00fd obr\u00e1zek jsem vygeneroval pomoc\u00ed aplikace Midjourney. \u010c\u00edm v\u00edce jsem se na n\u011bj d\u00edval, t\u00edm v\u00edce m\u011b zneklid\u0148oval. Estetika je temn\u00e1 a stra\u0161ideln\u00e1. Zd\u00e1 se, \u017ee zpoza objektu vystupuje st\u00ednov\u00e1 p\u0159\u00edtomnost. K vytvo\u0159en\u00ed tohoto obr\u00e1zku jsem pou\u017eil funkci estetick\u00e9ho odkazu aplikace Midjourney a jednodu\u0161e jsem mu zadal, aby to byla \"d\u00edvka stoj\u00edc\u00ed v klubu a usm\u00edvaj\u00edc\u00ed se. Uk\u00e1zal jsem obr\u00e1zek n\u011bkolika lidem a v\u0161ichni si mysleli, \u017ee je skute\u010dn\u00fd. <br><br>V\u011bt\u0161ina detektor\u016f ozna\u010dila obr\u00e1zek jako vytvo\u0159en\u00fd um\u011blou inteligenc\u00ed. Programy TruthScan a Sight Engine jej ozna\u010dily sk\u00f3re AI 99%, AI or Not 89%. <\/p>\n\n\n\n<p style=\"font-size:17px\">WasItAI si nebyl tak jist\u00fd jako detektory a klasifikoval sn\u00edmek pouze s v\u00fdsledkem 70% AI. <\/p>\n\n\n\n<p style=\"font-size:17px\">Detektor Winston AI naprosto selhal a ud\u011blil sn\u00edmku sk\u00f3re 1,56% AI. <\/p>\n\n\n\n<h2 class=\"wp-block-heading\">#10 \"Ice Box\"<strong> <\/strong>(vygeneroval Midjourney)<\/h2>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"771\" src=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/MJ-ICE-1024x771.jpg\" alt=\"Rozmazan\u00e1 fotografie, na n\u00ed\u017e je mu\u017e v \u010derven\u00e9 ko\u0161ili, kter\u00fd stoj\u00ed p\u0159ed n\u011bk\u00fdm.\" class=\"wp-image-5311\" title=\"\" srcset=\"https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/MJ-ICE-1024x771.jpg 1024w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/MJ-ICE-300x226.jpg 300w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/MJ-ICE-768x578.jpg 768w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/MJ-ICE-16x12.jpg 16w, https:\/\/research.undetectable.ai\/wp-content\/uploads\/2026\/02\/MJ-ICE.jpg 1232w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<style>\n @import url('https:\/\/fonts.googleapis.com\/css2?family=DM+Sans:wght@400;500;600;700&family=Outfit:wght@600;700;800&display=swap');\n\n .detector-scores-wrap {\n max-width: 780px;\n margin: 2rem auto;\n font-family: 'DM Sans', sans-serif;\n -webkit-font-smoothing: antialiased;\n }\n\n .detector-scores-wrap h3 {\n font-family: 'Outfit', sans-serif;\n font-size: 1.4rem;\n font-weight: 800;\n color: #1a1a2e;\n margin-bottom: 1.25rem;\n letter-spacing: -0.02em;\n }\n\n .detector-scores-table {\n width: 100%;\n border-collapse: separate;\n border-spacing: 0;\n border-radius: 14px;\n overflow: hidden;\n table-layout: fixed;\n box-shadow:\n 0 1px 3px rgba(0,0,0,0.06),\n 0 8px 24px rgba(0,0,0,0.08),\n 0 0 0 1px rgba(0,0,0,0.04);\n }\n\n .detector-scores-table thead th {\n background: #1a1a2e;\n color: #ffffff;\n font-family: 'Outfit', sans-serif;\n font-weight: 700;\n font-size: 0.7rem;\n text-transform: uppercase;\n letter-spacing: 0.04em;\n padding: 14px 6px;\n text-align: center;\n border: none;\n }\n\n .detector-scores-table thead th:first-child {\n text-align: left;\n padding-left: 14px;\n width: 18%;\n }\n\n .detector-scores-table thead th:nth-child(2) {\n width: 14%;\n }\n\n .detector-scores-table tbody tr {\n transition: background 0.2s ease;\n }\n\n .detector-scores-table tbody tr:hover {\n background: rgba(0,0,0,0.02);\n }\n\n .detector-scores-table tbody td {\n padding: 14px 6px;\n text-align: center;\n font-size: 0.9rem;\n font-weight: 500;\n color: #2d2d3a;\n border-bottom: 1px solid #eee;\n }\n\n .detector-scores-table tbody tr:last-child td {\n border-bottom: none;\n }\n\n .detector-scores-table tbody td:first-child {\n text-align: left;\n padding-left: 14px;\n font-weight: 600;\n color: #1a1a2e;\n }\n\n .score-pill {\n font-weight: 700;\n font-family: 'Outfit', sans-serif;\n font-size: 0.78rem;\n border-radius: 6px;\n padding: 5px 8px;\n display: inline-block;\n min-width: 0;\n text-align: center;\n white-space: nowrap;\n }\n\n .score-pass {\n background: #22c55e;\n color: #fff;\n }\n\n .score-warn {\n background: #fb923c;\n color: #fff;\n }\n\n .score-fail {\n background: #ef4444;\n color: #fff;\n }\n\n @media (max-width: 600px) {\n .detector-scores-table thead th,\n .detector-scores-table tbody td {\n padding: 10px 4px;\n font-size: 0.7rem;\n }\n .score-pill {\n padding: 4px 5px;\n font-size: 0.68rem;\n }\n .detector-scores-wrap h3 {\n font-size: 1.15rem;\n }\n }\n<\/style>\n\n<div class=\"detector-scores-wrap\">\n <h3>Jak to detektory obrazu vyhodnotily<\/h3>\n\n <table class=\"detector-scores-table\">\n <thead>\n <tr>\n <th>Obr\u00e1zek Model<\/th>\n <th>Kategorie<\/th>\n <th>TruthScan<\/th>\n <th>Sight Engine<\/th>\n <th>Um\u011bl\u00e1 inteligence nebo ne<\/th>\n <th>Winston AI<\/th>\n <th>WasItAI<\/th>\n <\/tr>\n <\/thead>\n <tbody>\n <tr>\n <td>MidJourney <span style=\"font-size:0.75rem; color:#888; display:block;\">Pln\u00e1 generace<\/span><\/td>\n <td style=\"font-size:0.82rem;\">Disinfo<\/td>\n <td><span class=\"score-pill score-pass\">99.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">92.00% AI<\/span><\/td>\n <td><span class=\"score-pill score-pass\">97.94% AI<\/span><\/td>\n <td><span class=\"score-pill score-fail\">41.20% AI<\/span><\/td>\n <td><span class=\"score-pill score-warn\">72.00% AI<\/span><\/td>\n <\/tr>\n <\/tbody>\n <\/table>\n<\/div>\n\n\n\n<p style=\"font-size:17px\">Tento obr\u00e1zek jsem vytvo\u0159il stejnou metodou uprost\u0159ed cesty jako v p\u0159edchoz\u00edm testu. Tentokr\u00e1t bylo c\u00edlem vytvo\u0159it n\u011bco trochu tmav\u0161\u00edho. Nejp\u0159esn\u011bj\u0161\u00edm detektorem byl TruthScan, kter\u00fd sn\u00edmek ozna\u010dil jako 99% AI. Programy Sight Engine a AI or Not oba spr\u00e1vn\u011b detekovaly obr\u00e1zek. Winston AI zcela selhal a WasItAI m\u011bl mnohem ni\u017e\u0161\u00ed sk\u00f3re spolehlivosti ne\u017e TruthScan, Sight Engine a AI or Not.<\/p>\n\n\n\n<p style=\"font-size:17px\">M\u011bl bych \u0159\u00edct, \u017ee v tomto p\u0159\u00edpad\u011b by se mnou vytvo\u0159en\u00fd obr\u00e1zek dal pou\u017e\u00edt nebo zar\u00e1movat dv\u011bma zp\u016fsoby. Na jedn\u00e9 stran\u011b by n\u011bkdo mohl pou\u017e\u00edt tento (nebo jak\u00fdkoli jin\u00fd typ m\u00e9dia s politickou tematikou\/ob\u010dansk\u00fdmi nepokoji) v n\u011bjak\u00e9m um\u011bleck\u00e9m nebo tv\u016fr\u010d\u00edm projektu, aby u\u010dinil prohl\u00e1\u0161en\u00ed. V takov\u00e9m p\u0159\u00edpad\u011b se riziko zd\u00e1 b\u00fdt ni\u017e\u0161\u00ed. Hlavn\u00ed obava, kterou m\u00e1m v souvislosti s touto kategori\u00ed m\u00e9di\u00ed generovan\u00fdch um\u011blou inteligenc\u00ed, spo\u010d\u00edv\u00e1 v tom, \u017ee je pou\u017e\u00edvaj\u00ed nekal\u00e9 osoby, kter\u00e9 tvrd\u00ed, \u017ee jsou skute\u010dn\u00e9.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Rekapitulace a v\u00fdsledky: Nejlep\u0161\u00ed detektory obrazu s um\u011blou inteligenc\u00ed<\/h2>\n\n\n\n<p style=\"font-size:17px\">Dob\u0159e, tak\u017ee te\u010f, kdy\u017e jsme zobrazili v\u0161echny testy, si je shrneme: Vygeneroval jsem 10 obr\u00e1zk\u016f um\u011bl\u00e9 inteligence<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li style=\"font-size:17px\">Pou\u017eil jsem ChatGPT, Nano Banana a Midjourney k vygenerov\u00e1n\u00ed 10 obr\u00e1zk\u016f AI.<\/li>\n\n\n\n<li style=\"font-size:17px\">Otestoval jsem p\u011bt detektor\u016f obr\u00e1zk\u016f AI tak, \u017ee jsem jimi nechal proj\u00edt v\u0161echny obr\u00e1zky AI, kter\u00e9 jsem vygeneroval.<\/li>\n\n\n\n<li style=\"font-size:17px\">Detektor TruthScan pro\u0161el v\u0161emi testy a byl nejp\u0159esn\u011bj\u0161\u00edm detektorem. AI or Not pro\u0161el 8 z 10 test\u016f, co\u017e ukazuje na ur\u010ditou spolehlivost. Sight Engine nepro\u0161el 3 z 10 test\u016f a prok\u00e1zal celkov\u011b spornou p\u0159esnost. Was It AI neusp\u011bl ve 4 z 10 test\u016f a jeho p\u0159esnost byla celkov\u011b n\u00edzk\u00e1. Winston AI byl nejm\u00e9n\u011b p\u0159esn\u00fdm detektorem obr\u00e1zk\u016f AI, pro\u0161el pouze 3 z 10 test\u016f a neust\u00e1le \u0161patn\u011b klasifikoval obr\u00e1zky.<\/li>\n<\/ul>\n\n\n\n<!DOCTYPE html>\n<html lang=\"en\">\n<head>\n<meta charset=\"UTF-8\">\n<meta name=\"viewport\" content=\"width=device-width, initial-scale=1.0\">\n<title>AI Image Detector - komplexn\u00ed v\u00fdsledky test\u016f<\/title>\n<style>\n @import url('https:\/\/fonts.googleapis.com\/css2?family=DM+Sans:wght@400;500;600;700&family=Outfit:wght@600;700;800&display=swap');\n\n *,*::before,*::after{box-sizing:border-box;margin:0;padding:0}\n\n body {\n background: #f4f3f0;\n min-height: 100vh;\n font-family: 'DM Sans', sans-serif;\n -webkit-font-smoothing: antialiased;\n color: #2d2d3a;\n }\n\n \/* \u2500\u2500 wrapper \u2500\u2500 *\/\n .results-page {\n max-width: 860px;\n margin: 0 auto;\n padding: 3rem 1.25rem 4rem;\n }\n\n \/* \u2500\u2500 hero header \u2500\u2500 *\/\n .hero {\n text-align: center;\n margin-bottom: 2.75rem;\n }\n .hero-badge {\n display: inline-block;\n font-family: 'Outfit', sans-serif;\n font-weight: 700;\n font-size: 0.65rem;\n text-transform: uppercase;\n letter-spacing: 0.12em;\n color: #fff;\n background: #1a1a2e;\n padding: 6px 16px;\n border-radius: 100px;\n margin-bottom: 1.1rem;\n }\n .hero h1 {\n font-family: 'Outfit', sans-serif;\n font-weight: 800;\n font-size: 2.4rem;\n color: #1a1a2e;\n letter-spacing: -0.03em;\n line-height: 1.15;\n margin-bottom: 0.6rem;\n }\n .hero p {\n font-size: 1.02rem;\n color: #6b6b7b;\n max-width: 560px;\n margin: 0 auto;\n line-height: 1.55;\n }\n\n \/* \u2500\u2500 accuracy cards row \u2500\u2500 *\/\n .accuracy-cards {\n display: grid;\n grid-template-columns: repeat(5, 1fr);\n gap: 10px;\n margin-bottom: 2.75rem;\n }\n .acc-card {\n background: #fff;\n border-radius: 14px;\n padding: 1.15rem 0.6rem 1rem;\n text-align: center;\n box-shadow:\n 0 1px 3px rgba(0,0,0,0.05),\n 0 6px 20px rgba(0,0,0,0.07),\n 0 0 0 1px rgba(0,0,0,0.03);\n position: relative;\n overflow: hidden;\n transition: transform 0.2s ease, box-shadow 0.2s ease;\n }\n .acc-card:hover {\n transform: translateY(-3px);\n box-shadow:\n 0 2px 6px rgba(0,0,0,0.06),\n 0 12px 32px rgba(0,0,0,0.1),\n 0 0 0 1px rgba(0,0,0,0.04);\n }\n .acc-card::before {\n content: '';\n position: absolute;\n top: 0; left: 0; right: 0;\n height: 4px;\n }\n .acc-card.rank-1::before { background: #16a34a; }\n .acc-card.rank-2::before { background: #a3e635; }\n .acc-card.rank-3::before { background: #facc15; }\n .acc-card.rank-4::before { background: #fb923c; }\n .acc-card.rank-5::before { background: #ef4444; }\n\n .acc-card .rank-num {\n font-family: 'Outfit', sans-serif;\n font-weight: 800;\n font-size: 0.62rem;\n text-transform: uppercase;\n letter-spacing: 0.08em;\n color: #999;\n margin-bottom: 0.35rem;\n }\n .acc-card .tool-name {\n font-family: 'Outfit', sans-serif;\n font-weight: 700;\n font-size: 0.92rem;\n color: #1a1a2e;\n margin-bottom: 0.65rem;\n line-height: 1.2;\n }\n .acc-ring {\n width: 72px; height: 72px;\n margin: 0 auto 0.55rem;\n position: relative;\n }\n .acc-ring svg { transform: rotate(-90deg); }\n .acc-ring .ring-bg { fill: none; stroke: #eee; stroke-width: 5; }\n .acc-ring .ring-fg { fill: none; stroke-width: 5; stroke-linecap: round; transition: stroke-dashoffset 1s ease; }\n .acc-ring .ring-label {\n position: absolute;\n inset: 0;\n display: flex;\n align-items: center;\n justify-content: center;\n font-family: 'Outfit', sans-serif;\n font-weight: 800;\n font-size: 1.1rem;\n color: #1a1a2e;\n }\n .acc-card .acc-fraction {\n font-size: 0.78rem;\n color: #888;\n font-weight: 500;\n }\n\n \/* \u2500\u2500 section titles \u2500\u2500 *\/\n .section-title {\n font-family: 'Outfit', sans-serif;\n font-weight: 800;\n font-size: 1.45rem;\n color: #1a1a2e;\n letter-spacing: -0.02em;\n margin-bottom: 1rem;\n }\n .section-subtitle {\n font-size: 0.88rem;\n color: #888;\n margin-bottom: 1.25rem;\n line-height: 1.5;\n }\n\n \/* \u2500\u2500 master table \u2500\u2500 *\/\n .master-table-wrap {\n border-radius: 14px;\n overflow: hidden;\n box-shadow:\n 0 1px 3px rgba(0,0,0,0.06),\n 0 8px 24px rgba(0,0,0,0.08),\n 0 0 0 1px rgba(0,0,0,0.04);\n margin-bottom: 2.75rem;\n background: #fff;\n }\n .master-table {\n width: 100%;\n border-collapse: collapse;\n table-layout: fixed;\n }\n .master-table thead th {\n background: #1a1a2e;\n color: #fff;\n font-family: 'Outfit', sans-serif;\n font-weight: 700;\n font-size: 0.62rem;\n text-transform: uppercase;\n letter-spacing: 0.05em;\n padding: 13px 5px;\n text-align: center;\n border: none;\n position: sticky;\n top: 0;\n z-index: 2;\n }\n .master-table thead th:first-child {\n text-align: left;\n padding-left: 14px;\n width: 14%;\n }\n .master-table thead th:nth-child(2) { width: 13%; }\n .master-table thead th:nth-child(3) { width: 9%; }\n\n .master-table tbody tr {\n transition: background 0.18s ease;\n }\n .master-table tbody tr:hover {\n background: rgba(26,26,46,0.025);\n }\n .master-table tbody td {\n padding: 11px 5px;\n text-align: center;\n font-size: 0.82rem;\n font-weight: 500;\n color: #2d2d3a;\n border-bottom: 1px solid #f0f0ec;\n }\n .master-table tbody tr:last-child td { border-bottom: none; }\n .master-table tbody td:first-child {\n text-align: left;\n padding-left: 14px;\n font-weight: 600;\n color: #1a1a2e;\n font-size: 0.78rem;\n line-height: 1.3;\n }\n .test-num {\n display: inline-block;\n font-family: 'Outfit', sans-serif;\n font-weight: 800;\n font-size: 0.6rem;\n background: #1a1a2e;\n color: #fff;\n width: 18px; height: 18px;\n line-height: 18px;\n text-align: center;\n border-radius: 5px;\n margin-right: 5px;\n vertical-align: middle;\n }\n .img-model-sub {\n display: block;\n font-size: 0.68rem;\n color: #999;\n font-weight: 400;\n margin-top: 1px;\n }\n .cat-badge {\n font-family: 'Outfit', sans-serif;\n font-size: 0.6rem;\n font-weight: 700;\n text-transform: uppercase;\n letter-spacing: 0.03em;\n padding: 3px 6px;\n border-radius: 5px;\n display: inline-block;\n white-space: normal;\n line-height: 1.35;\n max-width: 100%;\n }\n .cat-fraud { background: #fef2f2; color: #dc2626; }\n .cat-disinfo { background: #fefce8; color: #a16207; }\n .cat-general { background: #f0fdf4; color: #16a34a; }\n .cat-deepfake { background: #faf5ff; color: #7c3aed; }\n .cat-mixed { background: #fff7ed; color: #c2410c; }\n\n \/* score pills *\/\n .pill {\n font-weight: 700;\n font-family: 'Outfit', sans-serif;\n font-size: 0.7rem;\n border-radius: 6px;\n padding: 4px 7px;\n display: inline-block;\n text-align: center;\n white-space: nowrap;\n min-width: 68px;\n }\n .pill-pass { background: #22c55e; color: #fff; }\n .pill-warn { background: #fb923c; color: #fff; }\n .pill-fail { background: #ef4444; color: #fff; }\n\n \/* \u2500\u2500 legend \u2500\u2500 *\/\n .legend {\n display: flex;\n gap: 1.25rem;\n align-items: center;\n margin-bottom: 1.25rem;\n flex-wrap: wrap;\n }\n .legend-item {\n display: flex;\n align-items: center;\n gap: 6px;\n font-size: 0.78rem;\n font-weight: 500;\n color: #666;\n }\n .legend-dot {\n width: 12px; height: 12px;\n border-radius: 4px;\n }\n .legend-dot.lg-pass { background: #22c55e; }\n .legend-dot.lg-warn { background: #fb923c; }\n .legend-dot.lg-fail { background: #ef4444; }\n\n \/* \u2500\u2500 bottom summary strip \u2500\u2500 *\/\n .summary-strip {\n display: grid;\n grid-template-columns: repeat(3, 1fr);\n gap: 10px;\n margin-bottom: 2.5rem;\n }\n .strip-card {\n background: #1a1a2e;\n border-radius: 12px;\n padding: 1.1rem 1rem;\n color: #fff;\n text-align: center;\n }\n .strip-card .strip-val {\n font-family: 'Outfit', sans-serif;\n font-weight: 800;\n font-size: 1.65rem;\n letter-spacing: -0.02em;\n margin-bottom: 0.15rem;\n }\n .strip-card .strip-label {\n font-size: 0.72rem;\n color: rgba(255,255,255,0.55);\n font-weight: 500;\n text-transform: uppercase;\n letter-spacing: 0.05em;\n }\n .strip-val .hl-green { color: #4ade80; }\n .strip-val .hl-red { color: #f87171; }\n .strip-val .hl-amber { color: #fbbf24; }\n\n \/* \u2500\u2500 footer note \u2500\u2500 *\/\n .foot-note {\n font-size: 0.78rem;\n color: #999;\n line-height: 1.6;\n text-align: center;\n max-width: 640px;\n margin: 0 auto;\n }\n .foot-note strong { color: #666; font-weight: 600; }\n\n \/* \u2500\u2500 responsive \u2500\u2500 *\/\n @media (max-width: 700px) {\n .hero h1 { font-size: 1.6rem; }\n .accuracy-cards { grid-template-columns: repeat(2, 1fr); }\n .accuracy-cards .acc-card:last-child { grid-column: span 2; justify-self: center; max-width: 200px; }\n .summary-strip { grid-template-columns: 1fr; }\n .master-table thead th,\n .master-table tbody td { padding: 9px 3px; font-size: 0.68rem; }\n .pill { font-size: 0.6rem; padding: 3px 4px; min-width: 54px; }\n .test-num { width: 15px; height: 15px; line-height: 15px; font-size: 0.52rem; }\n .section-title { font-size: 1.15rem; }\n }\n\n \/* \u2500\u2500 entrance animations \u2500\u2500 *\/\n @keyframes fadeUp {\n from { opacity: 0; transform: translateY(18px); }\n to { opacity: 1; transform: translateY(0); }\n }\n .anim { animation: fadeUp 0.6s ease both; }\n .anim-d1 { animation-delay: 0.08s; }\n .anim-d2 { animation-delay: 0.16s; }\n .anim-d3 { animation-delay: 0.24s; }\n .anim-d4 { animation-delay: 0.32s; }\n .anim-d5 { animation-delay: 0.4s; }\n<\/style>\n<\/head>\n<body>\n\n<div class=\"results-page\">\n\n <!-- \u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550 HERO \u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550 -->\n <header class=\"hero anim\">\n <span class=\"hero-badge\">50 kontrol detekce - 5 detektor\u016f - 10 obr\u00e1zk\u016f AI<\/span>\n <h1>Komplexn\u00ed v\u00fdsledky test\u016f<\/h1>\n <p>Jak si vedlo p\u011bt popul\u00e1rn\u00edch detektor\u016f obr\u00e1zk\u016f s um\u011blou inteligenc\u00ed v oblasti podvod\u016f, dezinformac\u00ed, deepfakes a obecn\u00fdch fotografi\u00ed.<\/p>\n <\/header>\n\n <!-- \u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550 ACCURACY RING CARDS \u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550 -->\n <div class=\"accuracy-cards anim anim-d1\">\n\n <!-- 1 \u2014 TruthScan 100% -->\n <div class=\"acc-card rank-1\">\n <div class=\"rank-num\">#1<\/div>\n <div class=\"tool-name\">TruthScan<\/div>\n <div class=\"acc-ring\">\n <svg viewbox=\"0 0 80 80\" width=\"72\" height=\"72\">\n <circle class=\"ring-bg\" cx=\"40\" cy=\"40\" r=\"34\"\/>\n <circle class=\"ring-fg\" cx=\"40\" cy=\"40\" r=\"34\"\n stroke=\"#16a34a\"\n stroke-dasharray=\"213.6\"\n stroke-dashoffset=\"0\"\/>\n <\/svg>\n <span class=\"ring-label\">100%<\/span>\n <\/div>\n <div class=\"acc-fraction\">10 \/ 10 zji\u0161t\u011bno<\/div>\n <\/div>\n\n <!-- 2 \u2014 AI or Not 80% -->\n <div class=\"acc-card rank-2\">\n <div class=\"rank-num\">#2<\/div>\n <div class=\"tool-name\">Um\u011bl\u00e1 inteligence nebo ne<\/div>\n <div class=\"acc-ring\">\n <svg viewbox=\"0 0 80 80\" width=\"72\" height=\"72\">\n <circle class=\"ring-bg\" cx=\"40\" cy=\"40\" r=\"34\"\/>\n <circle class=\"ring-fg\" cx=\"40\" cy=\"40\" r=\"34\"\n stroke=\"#a3e635\"\n stroke-dasharray=\"213.6\"\n stroke-dashoffset=\"42.72\"\/>\n <\/svg>\n <span class=\"ring-label\">80%<\/span>\n <\/div>\n <div class=\"acc-fraction\">8 \/ 10 zji\u0161t\u011bno<\/div>\n <\/div>\n\n <!-- 3 \u2014 Sight Engine 70% -->\n <div class=\"acc-card rank-3\">\n <div class=\"rank-num\">#3<\/div>\n <div class=\"tool-name\">Sight Engine<\/div>\n <div class=\"acc-ring\">\n <svg viewbox=\"0 0 80 80\" width=\"72\" height=\"72\">\n <circle class=\"ring-bg\" cx=\"40\" cy=\"40\" r=\"34\"\/>\n <circle class=\"ring-fg\" cx=\"40\" cy=\"40\" r=\"34\"\n stroke=\"#facc15\"\n stroke-dasharray=\"213.6\"\n stroke-dashoffset=\"64.08\"\/>\n <\/svg>\n <span class=\"ring-label\">70%<\/span>\n <\/div>\n <div class=\"acc-fraction\">7 \/ 10 zji\u0161t\u011bno<\/div>\n <\/div>\n\n <!-- 4 \u2014 WasItAI 60% -->\n <div class=\"acc-card rank-4\">\n <div class=\"rank-num\">#4<\/div>\n <div class=\"tool-name\">WasItAI<\/div>\n <div class=\"acc-ring\">\n <svg viewbox=\"0 0 80 80\" width=\"72\" height=\"72\">\n <circle class=\"ring-bg\" cx=\"40\" cy=\"40\" r=\"34\"\/>\n <circle class=\"ring-fg\" cx=\"40\" cy=\"40\" r=\"34\"\n stroke=\"#fb923c\"\n stroke-dasharray=\"213.6\"\n stroke-dashoffset=\"85.44\"\/>\n <\/svg>\n <span class=\"ring-label\">60%<\/span>\n <\/div>\n <div class=\"acc-fraction\">6 \/ 10 zji\u0161t\u011bno<\/div>\n <\/div>\n\n <!-- 5 \u2014 Winston AI 30% -->\n <div class=\"acc-card rank-5\">\n <div class=\"rank-num\">#5<\/div>\n <div class=\"tool-name\">Winston AI<\/div>\n <div class=\"acc-ring\">\n <svg viewbox=\"0 0 80 80\" width=\"72\" height=\"72\">\n <circle class=\"ring-bg\" cx=\"40\" cy=\"40\" r=\"34\"\/>\n <circle class=\"ring-fg\" cx=\"40\" cy=\"40\" r=\"34\"\n stroke=\"#ef4444\"\n stroke-dasharray=\"213.6\"\n stroke-dashoffset=\"149.52\"\/>\n <\/svg>\n <span class=\"ring-label\">30%<\/span>\n <\/div>\n <div class=\"acc-fraction\">3 \/ 10 zji\u0161t\u011bno<\/div>\n <\/div>\n\n <\/div>\n\n <!-- \u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550 STATS STRIP \u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550 -->\n <div class=\"summary-strip anim anim-d2\">\n <div class=\"strip-card\">\n <div class=\"strip-val\"><span class=\"hl-green\">34<\/span> \/ 50<\/div>\n <div class=\"strip-label\">Celkov\u00fd po\u010det spr\u00e1vn\u00fdch detekc\u00ed<\/div>\n <\/div>\n <div class=\"strip-card\">\n <div class=\"strip-val\"><span class=\"hl-red\">16<\/span> \/ 50<\/div>\n <div class=\"strip-label\">Celkov\u00fd po\u010det zme\u0161kan\u00fdch detekc\u00ed<\/div>\n <\/div>\n <div class=\"strip-card\">\n <div class=\"strip-val\"><span class=\"hl-amber\">68%<\/span><\/div>\n <div class=\"strip-label\">Pr\u016fm\u011brn\u00e1 p\u0159esnost u v\u0161ech<\/div>\n <\/div>\n <\/div>\n\n <!-- \u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550 LEGEND \u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550 -->\n <h3 class=\"section-title anim anim-d3\">V\u0161ech 10 test\u016f - \u00fapln\u00e9 rozd\u011blen\u00ed<\/h3>\n <p class=\"section-subtitle anim anim-d3\">Ka\u017ed\u00e9 sk\u00f3re p\u0159edstavuje hodnocen\u00ed d\u016fv\u011bryhodnosti detektoru AI. Hranice pro vyhov\u011bn\u00ed: \u2265 90% AI.<\/p>\n\n <div class=\"legend anim anim-d3\">\n <div class=\"legend-item\"><span class=\"legend-dot lg-pass\"><\/span> Pr\u016fchod (\u2265 90%)<\/div>\n <div class=\"legend-item\"><span class=\"legend-dot lg-warn\"><\/span> T\u00e9m\u011b\u0159 ne\u00fasp\u011b\u0161n\u00fd pokus (70-89%)<\/div>\n <div class=\"legend-item\"><span class=\"legend-dot lg-fail\"><\/span> Selh\u00e1n\u00ed (< 70%)<\/div>\n <\/div>\n\n <!-- \u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550 MASTER TABLE \u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550 -->\n <div class=\"master-table-wrap anim anim-d4\">\n <table class=\"master-table\">\n <thead>\n <tr>\n <th>Test<\/th>\n <th>Kategorie<\/th>\n <th>Zdroj:<\/th>\n <th>TruthScan<\/th>\n <th>Sight Engine<\/th>\n <th>Um\u011bl\u00e1 inteligence nebo ne<\/th>\n <th>Winston AI<\/th>\n <th>WasItAI<\/th>\n <\/tr>\n <\/thead>\n <tbody>\n\n <!-- #1 -->\n <tr>\n <td><span class=\"test-num\">1<\/span> Mu\u017e na hran\u011b<\/td>\n <td><span class=\"cat-badge cat-general\">Obecn\u00e9<\/span><\/td>\n <td><span class=\"img-model-sub\">ChatGPT<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.00%<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.00%<\/span><\/td>\n <td><span class=\"pill pill-warn\">78.00%<\/span><\/td>\n <td><span class=\"pill pill-fail\">0.98%<\/span><\/td>\n <td><span class=\"pill pill-fail\">1.00%<\/span><\/td>\n <\/tr>\n\n <!-- #2 -->\n <tr>\n <td><span class=\"test-num\">2<\/span> Fale\u0161n\u00e1 \u00fa\u010dtenka<\/td>\n <td><span class=\"cat-badge cat-fraud\">Podvody<\/span><\/td>\n <td><span class=\"img-model-sub\">ChatGPT<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.00%<\/span><\/td>\n <td><span class=\"pill pill-fail\">19.00%<\/span><\/td>\n <td><span class=\"pill pill-pass\">94.48%<\/span><\/td>\n <td><span class=\"pill pill-fail\">0.04%<\/span><\/td>\n <td><span class=\"pill pill-fail\">1.00%<\/span><\/td>\n <\/tr>\n\n <!-- #3 -->\n <tr>\n <td><span class=\"test-num\">3<\/span> Bal\u00ed\u010dek nebezpe\u010d\u00ed<\/td>\n <td><span class=\"cat-badge cat-fraud\">Podvody<\/span><\/td>\n <td><span class=\"img-model-sub\">Nano ban\u00e1n<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.00%<\/span><\/td>\n <td><span class=\"pill pill-pass\">98.00%<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.10%<\/span><\/td>\n <td><span class=\"pill pill-fail\">31.98%<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.00%<\/span><\/td>\n <\/tr>\n\n <!-- #4 -->\n <tr>\n <td><span class=\"test-num\">4<\/span> \u0160v\u00e1b\u00ed mou\u010dka<\/td>\n <td><span class=\"cat-badge cat-fraud\">Podvody<\/span><\/td>\n <td><span class=\"img-model-sub\">Nano ban\u00e1n<\/span><\/td>\n <td><span class=\"pill pill-pass\">97.15%<\/span><\/td>\n <td><span class=\"pill pill-fail\">18.00%<\/span><\/td>\n <td><span class=\"pill pill-warn\">85.87%<\/span><\/td>\n <td><span class=\"pill pill-warn\">72.29%<\/span><\/td>\n <td><span class=\"pill pill-fail\">1.00%<\/span><\/td>\n <\/tr>\n\n <!-- #5 -->\n <tr>\n <td><span class=\"test-num\">5<\/span> Autovandalismus<\/td>\n <td><span class=\"cat-badge cat-mixed\">Podvody \/ dezinformace<\/span><\/td>\n <td><span class=\"img-model-sub\">Nano ban\u00e1n<\/span><\/td>\n <td><span class=\"pill pill-pass\">97.48%<\/span><\/td>\n <td><span class=\"pill pill-fail\">18.00%<\/span><\/td>\n <td><span class=\"pill pill-warn\">89.52%<\/span><\/td>\n <td><span class=\"pill pill-fail\">0.02%<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.00%<\/span><\/td>\n <\/tr>\n\n <!-- #6 -->\n <tr>\n <td><span class=\"test-num\">6<\/span> Soda Musk & Don<\/td>\n <td><span class=\"cat-badge cat-disinfo\">Disinfo<\/span><\/td>\n <td><span class=\"img-model-sub\">Nano ban\u00e1n<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.00%<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.00%<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.00%<\/span><\/td>\n <td><span class=\"pill pill-fail\">0.24%<\/span><\/td>\n <td><span class=\"pill pill-warn\">86.00%<\/span><\/td>\n <\/tr>\n\n <!-- #7 -->\n <tr>\n <td><span class=\"test-num\">7<\/span> St\u00e1le na\u017eivu<\/td>\n <td><span class=\"cat-badge cat-deepfake\">Deepfake<\/span><\/td>\n <td><span class=\"img-model-sub\">Nano ban\u00e1n<\/span><\/td>\n <td><span class=\"pill pill-pass\">97.49%<\/span><\/td>\n <td><span class=\"pill pill-pass\">95.00%<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.27%<\/span><\/td>\n <td><span class=\"pill pill-fail\">0.15%<\/span><\/td>\n <td><span class=\"pill pill-fail\">1.00%<\/span><\/td>\n <\/tr>\n\n <!-- #8 -->\n <tr>\n <td><span class=\"test-num\">8<\/span> Zneu\u017eit\u00ed psa Ben<\/td>\n <td><span class=\"cat-badge cat-mixed\">Deepfake \/ Disinfo<\/span><\/td>\n <td><span class=\"img-model-sub\">Nano ban\u00e1n<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.00%<\/span><\/td>\n <td><span class=\"pill pill-pass\">97.00%<\/span><\/td>\n <td><span class=\"pill pill-pass\">98.88%<\/span><\/td>\n <td><span class=\"pill pill-fail\">5.74%<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.00%<\/span><\/td>\n <\/tr>\n\n <!-- #9 -->\n <tr>\n <td><span class=\"test-num\">9<\/span> Klub Void<\/td>\n <td><span class=\"cat-badge cat-general\">Obecn\u00e9<\/span><\/td>\n <td><span class=\"img-model-sub\">MidJourney<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.00%<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.00%<\/span><\/td>\n <td><span class=\"pill pill-warn\">89.21%<\/span><\/td>\n <td><span class=\"pill pill-fail\">1.56%<\/span><\/td>\n <td><span class=\"pill pill-warn\">70.00%<\/span><\/td>\n <\/tr>\n\n <!-- #10 -->\n <tr>\n <td><span class=\"test-num\">10<\/span> Box na led<\/td>\n <td><span class=\"cat-badge cat-disinfo\">Disinfo<\/span><\/td>\n <td><span class=\"img-model-sub\">MidJourney<\/span><\/td>\n <td><span class=\"pill pill-pass\">99.00%<\/span><\/td>\n <td><span class=\"pill pill-pass\">92.00%<\/span><\/td>\n <td><span class=\"pill pill-pass\">97.94%<\/span><\/td>\n <td><span class=\"pill pill-fail\">41.20%<\/span><\/td>\n <td><span class=\"pill pill-warn\">72.00%<\/span><\/td>\n <\/tr>\n\n <\/tbody>\n <\/table>\n <\/div>\n\n <!-- \u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550 FOOTNOTES \u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550 -->\n <p class=\"foot-note anim anim-d5\">\n <strong>Metodika:<\/strong> Ka\u017ed\u00fd sn\u00edmek byl odesl\u00e1n jednou do ka\u017ed\u00e9ho detektoru. Sk\u00f3re <strong>\u2265 90% AI<\/strong> se pova\u017euje za spr\u00e1vnou detekci. V\u00fdsledky v rozmez\u00ed 70-89% jsou \"t\u00e9m\u011b\u0159 chybn\u00e9\". Cokoli pod 70% je ne\u00fasp\u011b\u0161n\u00e9. AI or Not ozna\u010dila jednu generaci na 78%, jednu na 85% a jednu na 89% - 89% je v souhrnn\u00e9 tabulce pova\u017eov\u00e1no za t\u00e9m\u011b\u0159 chybn\u00e9 s hv\u011bzdi\u010dkou.\n <\/p>\n\n<\/div>\n\n<\/body>\n<\/html>\n\n\n\n<h2 class=\"wp-block-heading\">Z\u00e1v\u011bre\u010dn\u00e9 pozn\u00e1mky a \u00fadaje<\/h2>\n\n\n\n<p style=\"font-size:17px\">Celkem jsem p\u0159i psan\u00ed tohoto \u010dl\u00e1nku a d\u016fkladn\u00e9m testov\u00e1n\u00ed str\u00e1vil 25 hodin. Multimod\u00e1ln\u00ed n\u00e1stroje pro detekci UI se st\u00e1le vyv\u00edjej\u00ed, ale je jasn\u00e9, \u017ee n\u011bkter\u00e9 jsou p\u0159esn\u011bj\u0161\u00ed ne\u017e jin\u00e9. Po otestov\u00e1n\u00ed v\u0161ech n\u00e1stroj\u016f je nejp\u0159esn\u011bj\u0161\u00ed n\u00e1stroj TruthScan. <a href=\"https:\/\/truthscan.com\/ai-image-detector\" target=\"_blank\" rel=\"noopener\">Detektor obrazu AI<\/a>. Testy mluv\u00ed samy za sebe. <\/p>\n\n\n\n<p style=\"font-size:17px\">Pokud byste cht\u011bli z\u00edskat p\u0159\u00edstup ke kopii CSV dat z testov\u00e1n\u00ed, kter\u00e9 jsem provedl v tomto \u010dl\u00e1nku, m\u016f\u017eete ji nal\u00e9zt <a href=\"https:\/\/docs.google.com\/spreadsheets\/d\/1iuapCicNcdG6dZCFd-dCmSRgUJ9GVCS2v9kh4PShRew\/edit?usp=sharing\" data-type=\"link\" data-id=\"https:\/\/docs.google.com\/spreadsheets\/d\/1iuapCicNcdG6dZCFd-dCmSRgUJ9GVCS2v9kh4PShRew\/edit?usp=sharing\" target=\"_blank\" rel=\"noopener\">zde.<\/a> Tabulka dat, na kterou jsem odk\u00e1zal, obsahuje v\u0161echny p\u016fvodn\u00ed v\u00fdzvy a v\u00fdsledky detekce z test\u016f uveden\u00fdch v tomto \u010dl\u00e1nku. <\/p>","protected":false},"excerpt":{"rendered":"<p>Pravd\u011bpodobn\u011b jste na internetu vid\u011bli v\u00edce fotografi\u00ed vytvo\u0159en\u00fdch um\u011blou inteligenc\u00ed, ne\u017e si mysl\u00edte. N\u011bkdy je z\u0159ejm\u00e9, \u017ee se jedn\u00e1 o [...]<\/p>","protected":false},"author":15,"featured_media":5326,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"rank_math_lock_modified_date":false,"_themeisle_gutenberg_block_has_review":false,"footnotes":""},"categories":[34],"tags":[19,96,97],"class_list":["post-5287","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-news","tag-ai-detection","tag-ai-images","tag-detecting-deepfakes"],"_links":{"self":[{"href":"https:\/\/research.undetectable.ai\/cs\/wp-json\/wp\/v2\/posts\/5287","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/research.undetectable.ai\/cs\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/research.undetectable.ai\/cs\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/research.undetectable.ai\/cs\/wp-json\/wp\/v2\/users\/15"}],"replies":[{"embeddable":true,"href":"https:\/\/research.undetectable.ai\/cs\/wp-json\/wp\/v2\/comments?post=5287"}],"version-history":[{"count":27,"href":"https:\/\/research.undetectable.ai\/cs\/wp-json\/wp\/v2\/posts\/5287\/revisions"}],"predecessor-version":[{"id":5334,"href":"https:\/\/research.undetectable.ai\/cs\/wp-json\/wp\/v2\/posts\/5287\/revisions\/5334"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/research.undetectable.ai\/cs\/wp-json\/wp\/v2\/media\/5326"}],"wp:attachment":[{"href":"https:\/\/research.undetectable.ai\/cs\/wp-json\/wp\/v2\/media?parent=5287"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/research.undetectable.ai\/cs\/wp-json\/wp\/v2\/categories?post=5287"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/research.undetectable.ai\/cs\/wp-json\/wp\/v2\/tags?post=5287"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}