{"id":161945,"date":"2026-02-21T17:11:54","date_gmt":"2026-02-21T15:11:55","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/oczenka-nadyozhnosti-parolej-sgenerirovannyh-bolshimi-yazykovymi-modelyami"},"modified":"2026-02-21T17:11:54","modified_gmt":"2026-02-21T15:11:55","slug":"oczenka-nadyozhnosti-parolej-sgenerirovannyh-bolshimi-yazykovymi-modelyami","status":"publish","type":"post","link":"https:\/\/prohoster.info\/da\/blog\/news\/oczenka-nadyozhnosti-parolej-sgenerirovannyh-bolshimi-yazykovymi-modelyami","title":{"rendered":"Vurdering af p\u00e5lideligheden af adgangskoder genereret af store sprogmodeller.","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

Resultaterne af analysen af sikkerheden ved adgangskoder genereret af store sprogmodeller og AI-assistenter er pr\u00e6senteret. Forskerne bad modellerne Claude, ChatGPT og Gemini om at generere en sikker 16-tegns adgangskode, og i alle tilf\u00e6lde fik de et resultat, der ved f\u00f8rste \u00f8jekast opfyldte alle krav til sikre adgangskoder og blev anerkendt af v\u00e6rkt\u00f8jer til kvalitetskontrol som sikker. Adgangskoderne indeholdt symboler i forskellige bogstavsfelter, specialtegn og tal, men de s\u00e5 kun sikre ud; faktisk havde de minimal entropi, blev dannet efter en typisk skabelon og dannede m\u00f8nstre ved gentagne foresp\u00f8rgsler. <\/p>\n

If\u00f8lge forskerne bruges de forudsigelige adgangskoder genereret af store sprogmodeller p\u00e5 praktisk h\u00e5ndtering af reelle brugere og foresl\u00e5s under kodningsarbejde af AI-assistenter. Entropiniveauet i de genererede adgangskoder fra AI-modellerne vurderes til 20-27 bit, hvilket kr\u00e6ver fra flere sekunder til timer for at kn\u00e6kke adgangskoden, mens analyseresultaterne fra kvalitetskontrolv\u00e6rkt\u00f8jer forudsiger tid til at kn\u00e6kke i flere \u00e5rhundreder. Skabelonen for s\u00e5danne adgangskoder er et resultat af opbygningen af indhold af store sprogmodeller baseret p\u00e5 forudsigelse af tokens. <\/p>\n

Af de 50 passwords genereret i Claude Opus 4.6, gentog 18 sig helt, alle passwords begyndte med et bogstav (prim\u00e6rt «G»), efterfulgt af et tal (prim\u00e6rt 7), og alle passwords indeholdt tegnene «L», «9», «m», «2», «$» og «#».

\"Vurdering<\/center> <\/p>\n

I GPT-5.2 begyndte n\u00e6sten alle passwords med bogstavet «v», efterfulgt af bogstavet «Q» i halvdelen af passwords og et gentagende m\u00f8nster af et begr\u00e6nset antal tegn. I Gemini 3 begyndte n\u00e6sten halvdelen af passwords med tegnene «K» eller «k», efterfulgt af «#», «P» eller «9», og det anvendte tegns\u00e6t var st\u00e6rkt begr\u00e6nset. En stigning i «temperaturen» ved foresp\u00f8rgsler til AI-modeller p\u00e5virker ikke m\u00e6rkbart kvaliteten af de genererede passwords. <\/p>\n

N\u00e5r det kommer til brugen af AI-assistenter, afh\u00e6nger kvaliteten af passwordet i h\u00f8j grad af den foresp\u00f8rgsel, som udvikleren har udformet. For eksempel genererede Claude Code med Opus 4.6 og Gemini-CLI med Auto Gemini 3 ved foresp\u00f8rgslen om at generere et sikkert password kommandoen «openssl rand». Her brugte Gemini-CLI med Auto Gemini 3 ved foresp\u00f8rgslen «generer et password» «openssl rand», mens det ved foresp\u00f8rgslen «foresl\u00e5 et password» genererede passwordet gennem AI-modellen. Codex med GPT-5.3-Code brugte til tider en ekstern utility til at generere et sikkert password, men genererede til tider forudsigelige passwords p\u00e5 egen h\u00e5nd. Claude Code med Opus 4.5 genererede oftest forudsigelige passwords selv. Browseren ChatGPT Atlas skabte et usikkert password ved hj\u00e6lp af AI-modellen, n\u00e5r den skulle finde p\u00e5 et password til registrering p\u00e5 en hjemmeside.
\n
Kilde: opennet.ru<\/a> <\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"

\u041f\u0440\u0435\u0434\u0441\u0442\u0430\u0432\u043b\u0435\u043d\u044b \u0440\u0435\u0437\u0443\u043b\u044c\u0442\u0430\u0442\u044b \u0430\u043d\u0430\u043b\u0438\u0437\u0430 \u043d\u0430\u0434\u0451\u0436\u043d\u043e\u0441\u0442\u0438 \u043f\u0430\u0440\u043e\u043b\u0435\u0439, \u0433\u0435\u043d\u0435\u0440\u0438\u0440\u0443\u0435\u043c\u044b\u0445 \u0431\u043e\u043b\u044c\u0448\u0438\u043c\u0438 \u044f\u0437\u044b\u043a\u043e\u0432\u044b\u043c\u0438 \u043c\u043e\u0434\u0435\u043b\u044f\u043c\u0438 \u0438 AI-\u0430\u0441\u0441\u0438\u0441\u0442\u0435\u043d\u0442\u0430\u043c\u0438. \u0418\u0441\u0441\u043b\u0435\u0434\u043e\u0432\u0430\u0442\u0435\u043b\u0438 \u043f\u043e\u043f\u0440\u043e\u0441\u0438\u043b\u0438 \u043c\u043e\u0434\u0435\u043b\u0438 Claude, ChatGPT \u0438 Gemini \u0441\u0433\u0435\u043d\u0435\u0440\u0438\u0440\u043e\u0432\u0430\u0442\u044c \u043d\u0430\u0434\u0451\u0436\u043d\u044b\u0439 16-\u0441\u0438\u043c\u0432\u043e\u043b\u044c\u043d\u044b\u0439 \u043f\u0430\u0440\u043e\u043b\u044c \u0438 \u0432\u043e \u0432\u0441\u0435\u0445 \u0441\u043b\u0443\u0447\u0430\u044f\u0445 \u043f\u043e\u043b\u0443\u0447\u0438\u043b\u0438 \u0440\u0435\u0437\u0443\u043b\u044c\u0442\u0430\u0442, \u043d\u0430 \u043f\u0435\u0440\u0432\u044b\u0439 \u0432\u0437\u0433\u043b\u044f\u0434 \u043e\u0442\u0432\u0435\u0447\u0430\u044e\u0449\u0438\u0439 \u0432\u0441\u0435\u043c \u0442\u0440\u0435\u0431\u043e\u0432\u0430\u043d\u0438\u044f\u043c \u043a \u0431\u0435\u0437\u043e\u043f\u0430\u0441\u043d\u044b\u043c \u043f\u0430\u0440\u043e\u043b\u044f\u043c \u0438 \u043f\u0440\u0438\u0437\u043d\u0430\u0432\u0430\u0435\u043c\u044b\u0439 \u0443\u0442\u0438\u043b\u0438\u0442\u0430\u043c\u0438 \u0434\u043b\u044f \u043f\u0440\u043e\u0432\u0435\u0440\u043a\u0438 \u043a\u0430\u0447\u0435\u0441\u0442\u0432\u0430 \u043f\u0430\u0440\u043e\u043b\u0435\u0439 \u043a\u0430\u043a \u043d\u0430\u0434\u0451\u0436\u043d\u044b\u0439. \u041f\u0430\u0440\u043e\u043b\u0438 \u0441\u043e\u0447\u0435\u0442\u0430\u043b\u0438 \u0432 \u0441\u0435\u0431\u0435 \u0441\u0438\u043c\u0432\u043e\u043b\u044b \u0432 \u0440\u0430\u0437\u043d\u044b\u0445 \u0440\u0435\u0433\u0438\u0441\u0442\u0440\u0430\u0445, \u0441\u043f\u0435\u0446\u0441\u0438\u043c\u0432\u043e\u043b\u044b […]<\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"author":8,"featured_media":161946,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[702],"tags":[],"class_list":["post-161945","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news"],"aioseo_notices":[],"aioseo_head":"\n\t\t\n\t\n\t\n\t\n\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\n","aioseo_head_json":{"title":"\ud83e\udd47Vurdering af sikkerheden ved kodeord genereret af store sprogmodeller | ProHoster","description":"Resultaterne af analysen af sikkerheden ved kodeord genereret af store sprogmodeller og AI-assistenter pr\u00e6senteres.","canonical_url":"https:\/\/prohoster.info\/da\/blog\/news\/oczenka-nadyozhnosti-parolej-sgenerirovannyh-bolshimi-yazykovymi-modelyami","robots":"max-image-preview:large","keywords":"","webmasterTools":{"miscellaneous":""},"schema":null,"og:locale":"da_DK","og:site_name":"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b","og:type":"article","og:title":"\ud83e\udd47\u041e\u0446\u0435\u043d\u043a\u0430 \u043d\u0430\u0434\u0451\u0436\u043d\u043e\u0441\u0442\u0438 \u043f\u0430\u0440\u043e\u043b\u0435\u0439, \u0441\u0433\u0435\u043d\u0435\u0440\u0438\u0440\u043e\u0432\u0430\u043d\u043d\u044b\u0445 \u0431\u043e\u043b\u044c\u0448\u0438\u043c\u0438 \u044f\u0437\u044b\u043a\u043e\u0432\u044b\u043c\u0438 \u043c\u043e\u0434\u0435\u043b\u044f\u043c\u0438 | ProHoster","og:description":"\u041f\u0440\u0435\u0434\u0441\u0442\u0430\u0432\u043b\u0435\u043d\u044b \u0440\u0435\u0437\u0443\u043b\u044c\u0442\u0430\u0442\u044b \u0430\u043d\u0430\u043b\u0438\u0437\u0430 \u043d\u0430\u0434\u0451\u0436\u043d\u043e\u0441\u0442\u0438 \u043f\u0430\u0440\u043e\u043b\u0435\u0439, \u0433\u0435\u043d\u0435\u0440\u0438\u0440\u0443\u0435\u043c\u044b\u0445 \u0431\u043e\u043b\u044c\u0448\u0438\u043c\u0438 \u044f\u0437\u044b\u043a\u043e\u0432\u044b\u043c\u0438 \u043c\u043e\u0434\u0435\u043b\u044f\u043c\u0438 \u0438 AI-\u0430\u0441\u0441\u0438\u0441\u0442\u0435\u043d\u0442\u0430\u043c\u0438.","og:url":"https:\/\/prohoster.info\/da\/blog\/news\/oczenka-nadyozhnosti-parolej-sgenerirovannyh-bolshimi-yazykovymi-modelyami","og:image":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:secure_url":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:width":350,"og:image:height":350,"article:published_time":"2026-02-21T15:11:55+00:00","article:modified_time":"2026-02-21T15:11:55+00:00","article:publisher":"https:\/\/www.facebook.com\/prohoster","article:author":"https:\/\/www.facebook.com\/prohoster"},"aioseo_meta_data":{"post_id":"161945","title":null,"description":null,"keywords":null,"keyphrases":null,"primary_term":null,"canonical_url":null,"og_title":null,"og_description":null,"og_object_type":"default","og_image_type":"default","og_image_url":null,"og_image_width":null,"og_image_height":null,"og_image_custom_url":null,"og_image_custom_fields":null,"og_video":null,"og_custom_url":null,"og_article_section":null,"og_article_tags":null,"twitter_use_og":false,"twitter_card":"default","twitter_image_type":"default","twitter_image_url":null,"twitter_image_custom_url":null,"twitter_image_custom_fields":null,"twitter_title":null,"twitter_description":null,"schema":{"blockGraphs":[],"customGraphs":[],"default":{"data":{"Article":[],"Course":[],"Dataset":[],"FAQPage":[],"Movie":[],"Person":[],"Product":[],"ProductReview":[],"Car":[],"Recipe":[],"Service":[],"SoftwareApplication":[],"WebPage":[]},"graphName":"","isEnabled":true},"graphs":[]},"schema_type":"default","schema_type_options":null,"pillar_content":false,"robots_default":true,"robots_noindex":false,"robots_noarchive":false,"robots_nosnippet":false,"robots_nofollow":false,"robots_noimageindex":false,"robots_noodp":false,"robots_notranslate":false,"robots_max_snippet":null,"robots_max_videopreview":null,"robots_max_imagepreview":"large","priority":null,"frequency":null,"local_seo":null,"seo_analyzer_scan_date":"2026-02-21 15:12:55","breadcrumb_settings":null,"limit_modified_date":false,"reviewed_by":null,"ai":null,"created":"2026-02-21 15:11:55","updated":"2026-02-21 15:12:55","focus_keyword":null,"additional_keywords":null,"truseo_locale":null},"gt_translate_keys":[{"key":"link","format":"url"}],"_links":{"self":[{"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/posts\/161945","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/comments?post=161945"}],"version-history":[{"count":0,"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/posts\/161945\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/media\/161946"}],"wp:attachment":[{"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/media?parent=161945"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/categories?post=161945"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/prohoster.info\/da\/wp-json\/wp\/v2\/tags?post=161945"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}