{"id":17467,"date":"2025-09-03T11:15:35","date_gmt":"2025-09-03T09:15:35","guid":{"rendered":"https:\/\/www.slovensko.ai\/?p=17467"},"modified":"2025-09-03T12:43:01","modified_gmt":"2025-09-03T10:43:01","slug":"chatgpt-zavadza-rodicovske-prvky-na-ochranu-tinedzerov-v-krize","status":"publish","type":"post","link":"https:\/\/www.slovensko.ai\/en\/chatgpt-zavadza-rodicovske-prvky-na-ochranu-tinedzerov-v-krize\/","title":{"rendered":"ChatGPT zav\u00e1dza rodi\u010dovsk\u00e9 prvky na ochranu t\u00edned\u017eerov v kr\u00edze"},"content":{"rendered":"
<\/span> 440<\/span> Viden\u00ed<\/span><\/div>

Spolo\u010dnos\u0165 OpenAI \u010del\u00ed vln\u00e1m kritiky po tragickom pr\u00edpade 16-ro\u010dn\u00e9ho t\u00edned\u017eera, ktor\u00fd si \u00fadajne vzal \u017eivot po rozhovoroch s chatbotom ChatGPT<\/a>. Rodi\u010dia zosnul\u00e9ho podali \u017ealobu, v ktorej tvrdia, \u017ee chatbot ich syna v jeho rozhodnut\u00ed podporoval, a dokonca mu poskytol n\u00e1vod na uskuto\u010dnenie samovra\u017edy.<\/p>\n

Na z\u00e1klade tohto pr\u00edpadu a rast\u00facich ob\u00e1v verejnosti OpenAI ozn\u00e1milo s\u00e9riu z\u00e1sadn\u00fdch zmien v pr\u00edstupe k bezpe\u010dnosti pou\u017e\u00edvate\u013eov, najm\u00e4 mladistv\u00fdch.<\/p>\n

Rodi\u010do<\/strong>vsk\u00e9 kontroly ako nov\u00fd \u0161tandard<\/strong><\/h2>\n

OpenAI pl\u00e1nuje do jedn\u00e9ho mesiaca zavies\u0165 rodi\u010dovsk\u00e9 kontroly v r\u00e1mci slu\u017eby ChatGPT<\/a>. Funkcia umo\u017en\u00ed rodi\u010dom prepoji\u0165 svoj \u00fa\u010det s \u00fa\u010dtom t\u00edned\u017eera prostredn\u00edctvom jednoduch\u00e9ho e-mailov\u00e9ho pozvania. N\u00e1sledne bud\u00fa ma\u0165 rodi\u010dia mo\u017enos\u0165 nastavi\u0165 parametre spr\u00e1vania modelu od vhodnosti odpoved\u00ed po \u00farove\u0148 ochrany citliv\u00fdch t\u00e9m.<\/p>\n

Tieto nastavenia bud\u00fa ma\u0165 predvolen\u00fa konfigur\u00e1ciu pod\u013ea veku die\u0165a\u0165a, \u010d\u00edm sa zabezpe\u010d\u00ed, \u017ee chatbot nebude zobrazova\u0165 ani generova\u0165 nevhodn\u00fd alebo potenci\u00e1lne \u0161kodliv\u00fd obsah.<\/p>\n

Sledovanie znakov emocion\u00e1lnej n\u00fadze<\/strong><\/h2>\n

Jedn\u00fdm z najz\u00e1sadnej\u0161\u00edch prvkov pripravovan\u00fdch opatren\u00ed je schopnos\u0165 syst\u00e9mu identifikova\u0165 zn\u00e1mky ak\u00fatnej emo\u010dnej n\u00fadze. V tak\u00fdchto pr\u00edpadoch bude ChatGPT schopn\u00fd:<\/p>\n

    \n
  • \n

    upozorni\u0165 rodi\u010dov na potenci\u00e1lne riziko<\/p>\n<\/li>\n

  • \n

    preru\u0161i\u0165 alebo obmedzi\u0165 konverz\u00e1ciu<\/p>\n<\/li>\n

  • \n

    presmerova\u0165 pou\u017e\u00edvate\u013ea na \u0161pecializovan\u00fd model so zv\u00fd\u0161enou mierou bezpe\u010dnosti<\/p>\n<\/li>\n<\/ul>\n

    Cie\u013eom je zabr\u00e1ni\u0165 tomu, aby chatbot sl\u00fa\u017eil ako \u00fanikov\u00fd priestor pre mladistv\u00fdch v kr\u00edzov\u00fdch situ\u00e1ci\u00e1ch bez mo\u017enosti vonkaj\u0161ej pomoci.<\/p>\n

    Deaktiv\u00e1cia pam\u00e4te a hist\u00f3rie pre zranite\u013en\u00fdch pou\u017e\u00edvate\u013eov<\/strong><\/h2>\n

    OpenAI tie\u017e umo\u017en\u00ed rodi\u010dom a samotn\u00fdm pou\u017e\u00edvate\u013eom vypn\u00fa\u0165 funkcie ako pam\u00e4\u0165 a hist\u00f3ria konverz\u00e1ci\u00ed, ktor\u00e9 m\u00f4\u017eu podporova\u0165 emocion\u00e1lnu z\u00e1vislos\u0165 od chatbota. V\u00fdvoj\u00e1ri upozor\u0148uj\u00fa, \u017ee pre niektor\u00fdch t\u00edned\u017eerov m\u00f4\u017ee AI predstavova\u0165 zdroj falo\u0161nej emocion\u00e1lnej stability, \u010do m\u00f4\u017ee zhor\u0161i\u0165 ich psychick\u00e9 zdravie.<\/p>\n

    Vypnut\u00edm t\u00fdchto funkci\u00ed sa z\u00e1rove\u0148 zn\u00ed\u017ei riziko opakovania toxick\u00fdch alebo nebezpe\u010dn\u00fdch vzorcov v interakci\u00e1ch.<\/p>\n

    Citliv\u00e9 rozhovory presmeruje robustnej\u0161\u00ed model<\/strong><\/h2>\n

    \u010eal\u0161ou d\u00f4le\u017eitou zmenou bude automatick\u00e9 presmerovanie citliv\u00fdch konverz\u00e1ci\u00ed, napr\u00edklad o samovra\u017ede, sebapo\u0161kodzovan\u00ed \u010di z\u00e1vislostiach, na robustnej\u0161ie modely, ktor\u00e9 preuk\u00e1zali vy\u0161\u0161iu konzistentnos\u0165 a schopnos\u0165 dodr\u017eiava\u0165 etick\u00e9 z\u00e1sady.<\/p>\n

    Ide o tzv. reasoning modely ako GPT-5 thinking \u010di modely s\u00e9rie o3, ktor\u00e9 bud\u00fa nastaven\u00e9 tak, aby sa vyhli zdie\u013eaniu ak\u00e9hoko\u013evek n\u00e1vodu alebo neprimeran\u00e9ho obsahu a z\u00e1rove\u0148 sa sna\u017eili vies\u0165 konverz\u00e1ciu k z\u00edskaniu pomoci.<\/p>\n

    Spolupr\u00e1ca s odborn\u00edkmi na du\u0161evn\u00e9 zdravie<\/strong><\/h2>\n

    OpenAI deklaruje, \u017ee tieto kroky s\u00fa s\u00fa\u010das\u0165ou \u0161ir\u0161ej bezpe\u010dnostnej strat\u00e9gie. Spolo\u010dnos\u0165 v s\u00fa\u010dasnosti spolupracuje s medzin\u00e1rodn\u00fdmi expertmi na du\u0161evn\u00e9 zdravie, v\u00fdvin det\u00ed a dizajn interakcie \u010dlovek\u2013stroj. Vznikla aj odborn\u00e1 skupina Expert Council on Well\u2011Being and AI, ktorej cie\u013eom je poskytova\u0165 odpor\u00fa\u010dania pri v\u00fdvoji bezpe\u010dn\u00fdch funkcional\u00edt.<\/p>\n

    Okrem toho OpenAI zapojilo medzin\u00e1rodn\u00fa sie\u0165 lek\u00e1rov a terapeutov, ktor\u00ed bud\u00fa pom\u00e1ha\u0165 nastavova\u0165 \u0161tandardy pre detekciu kr\u00edzov\u00e9ho spr\u00e1vania v komunik\u00e1cii.<\/p>\n

    Reakcie verejnosti<\/strong><\/h2>\n

    Mnoh\u00ed experti v\u00edtaj\u00fa nov\u00fd smer ako v\u00fdznamn\u00fd krok vpred, no z\u00e1rove\u0148 poukazuj\u00fa na to, \u017ee zmeny s\u00fa zatia\u013e vo f\u00e1ze pl\u00e1novania a nepon\u00fakaj\u00fa okam\u017eit\u00e9 rie\u0161enia. Kritici ozna\u010duj\u00fa ozn\u00e1menia za nejasn\u00e9 s\u013euby, ktor\u00e9 neprin\u00e1\u0161aj\u00fa okam\u017eit\u00fa ochranu pou\u017e\u00edvate\u013eov.<\/p>\n

    Rodiny obet\u00ed navy\u0161e \u017eiadaj\u00fa v\u00fdraznej\u0161ie obmedzenia v pr\u00edstupe mladistv\u00fdch k ve\u013ek\u00fdm jazykov\u00fdm modelom, alebo ich \u00fapln\u00e9 stiahnutie z trhu, k\u00fdm nebude dostupn\u00e1 \u00fa\u010dinn\u00e1 ochrana.<\/p>\n

    Z\u00e1ver: Bezpe\u010dnostn\u00e1 \u00e9ra umelej inteligencie pr\u00e1ve za\u010d\u00edna<\/strong><\/h2>\n

    Trag\u00e9die spojen\u00e9 s vyu\u017e\u00edvan\u00edm AI chatbotov medzi mladistv\u00fdmi jasne ukazuj\u00fa potrebu r\u00fdchlej a d\u00f4slednej reakcie v\u00fdvoj\u00e1rov. OpenAI svoj\u00edm pl\u00e1nom zavies\u0165 rodi\u010dovsk\u00e9 kontroly a syst\u00e9m varovan\u00ed nastavuje nov\u00fd \u0161tandard v oblasti zodpovedn\u00e9ho dizajnu umelej inteligencie.<\/p>\n

    Zost\u00e1va v\u0161ak otvorenou ot\u00e1zkou, \u010di a ako r\u00fdchlo bud\u00fa tieto opatrenia implementovan\u00e9 a ak\u00fd bud\u00fa ma\u0165 re\u00e1lny dopad. Jedno je v\u0161ak ist\u00e9 \u00e9ra regulovanej, citlivej a bezpe\u010dnej AI pr\u00e1ve za\u010d\u00edna.<\/p>","protected":false},"excerpt":{"rendered":"

    440 Viden\u00edSpolo\u010dnos\u0165 OpenAI \u010del\u00ed vln\u00e1m kritiky po tragickom pr\u00edpade 16-ro\u010dn\u00e9ho t\u00edned\u017eera, ktor\u00fd si \u00fadajne vzal \u017eivot po rozhovoroch s chatbotom ChatGPT. Rodi\u010dia zosnul\u00e9ho podali \u017ealobu, v ktorej tvrdia, \u017ee chatbot ich syna v jeho rozhodnut\u00ed podporoval, a dokonca mu poskytol n\u00e1vod na uskuto\u010dnenie samovra\u017edy. Na z\u00e1klade tohto pr\u00edpadu a rast\u00facich ob\u00e1v verejnosti OpenAI ozn\u00e1milo s\u00e9riu […]<\/p>\n","protected":false},"author":5,"featured_media":17468,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[51],"tags":[377,53],"class_list":["post-17467","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-blog","tag-ai-novinky","tag-umela-inteligencia"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.slovensko.ai\/en\/wp-json\/wp\/v2\/posts\/17467"}],"collection":[{"href":"https:\/\/www.slovensko.ai\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.slovensko.ai\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.slovensko.ai\/en\/wp-json\/wp\/v2\/users\/5"}],"replies":[{"embeddable":true,"href":"https:\/\/www.slovensko.ai\/en\/wp-json\/wp\/v2\/comments?post=17467"}],"version-history":[{"count":1,"href":"https:\/\/www.slovensko.ai\/en\/wp-json\/wp\/v2\/posts\/17467\/revisions"}],"predecessor-version":[{"id":17469,"href":"https:\/\/www.slovensko.ai\/en\/wp-json\/wp\/v2\/posts\/17467\/revisions\/17469"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.slovensko.ai\/en\/wp-json\/wp\/v2\/media\/17468"}],"wp:attachment":[{"href":"https:\/\/www.slovensko.ai\/en\/wp-json\/wp\/v2\/media?parent=17467"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.slovensko.ai\/en\/wp-json\/wp\/v2\/categories?post=17467"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.slovensko.ai\/en\/wp-json\/wp\/v2\/tags?post=17467"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}