Chatbot-i i inteligjenc\u00ebs artificiale, ChatGPT, ka marr\u00eb v\u00ebmendje p\u00ebr prirjen e tij p\u00ebr t\u00eb ofruar informacione t\u00eb pasakta. Megjith\u00ebse OpenAI paralajm\u00ebron n\u00eb faqen kryesore se AI mund t\u00eb “gjeneroj\u00eb her\u00eb pas here informacione t\u00eb pasakta”, shum\u00eb kritik\u00eb argumentojn\u00eb se ky paralajm\u00ebrim \u00ebsht\u00eb i pamjaftuesh\u00ebm, duke pasur parasysh pasojat e mundshme t\u00eb keqinformimit.<\/p>\n\n\n\n
Kufizimet e chatbot-it u vun\u00eb re kur avokati amerikan Steven A. Schwartz e p\u00ebrdori at\u00eb p\u00ebr t\u00eb gjetur \u00e7\u00ebshtje mb\u00ebshtet\u00ebse p\u00ebr nj\u00eb padi kund\u00ebr linj\u00ebs ajrore kolumbiane, Avianca. N\u00eb fakt, rastet e ofruara nga ChatGPT nuk ekzistonin. Pavar\u00ebsisht dyshimit t\u00eb tij fillestar dhe kontrolleve t\u00eb m\u00ebvonshme me chatbot-in, sistemi vazhdoi t\u00eb pohonte se rastet e fabrikuara ishin legjitime.<\/p>\n\n\n\n
Nd\u00ebrsa Schwartz mban nj\u00eb pjes\u00eb t\u00eb fajit, shpesht\u00ebsia me t\u00eb cil\u00ebn p\u00ebrdoruesit i marrin t\u00eb mir\u00ebq\u00ebna si t\u00eb besueshme pohimet e ChatGPT, k\u00ebrkon nj\u00eb konsiderat\u00eb m\u00eb t\u00eb gjer\u00eb. Gjat\u00eb muajve t\u00eb fundit, jan\u00eb shfaqur raporte t\u00eb shumta mbi njer\u00ebzit q\u00eb mashtrohen nga g\u00ebnjeshtrat e chatbot-it. Keqinformimi varionte nga fabrikimi i lajmeve dhe punimeve akademike, deri te gjenerimi i lajmeve t\u00eb pasakta ose p\u00ebrhapja e akuzave t\u00eb rreme p\u00ebr sjellje t\u00eb pahijshme seksuale t\u00eb personave publik\u00eb.<\/p>\n\n\n\n
K\u00ebto raste nxjerrin n\u00eb pah nj\u00eb potencial shqet\u00ebsues p\u00ebr pasoja t\u00eb d\u00ebmshme p\u00ebr shkak t\u00eb keqinformimit t\u00eb ChatGPT. Gjithashtu, k\u00ebto raste n\u00ebnvizojn\u00eb nevoj\u00ebn q\u00eb p\u00ebrdoruesit t’i qasen p\u00ebrgjigjeve t\u00eb gjeneruara nga AI me nj\u00eb sy kritik, dhe p\u00ebr m\u00eb tep\u00ebr \u00ebsht\u00eb e r\u00ebnd\u00ebsishme q\u00eb OpenAI t\u00eb p\u00ebrmir\u00ebsoj\u00eb besueshm\u00ebrin\u00eb e sistemit.<\/p>\n\n\n\n
Nd\u00ebrsa raporte t\u00eb ndryshme shpesh paraqesin chatbot-et AI, si ChatGPT, si t\u00eb ngjash\u00ebm me inteligjenc\u00ebn njer\u00ebzore, ata shpesh minimizojn\u00eb kufizimet dhe potencialin e robot\u00ebve p\u00ebr t\u00eb ofruar informacione t\u00eb pasakta. Sipas shum\u00eb ekspert\u00ebve, OpenAI duhet t\u00eb nd\u00ebrmarr\u00eb hapa m\u00eb t\u00eb m\u00ebdhenj p\u00ebr t\u00eb siguruar sakt\u00ebsin\u00eb e informacionit.<\/p>\n\n\n\n
Megjith\u00ebse chatbot-et po promovohen si nj\u00eb lloj i ri teknologjie, p\u00ebrdoruesit shpesh i trajtojn\u00eb ato si motorr\u00eb k\u00ebrkimi. Kjo \u00ebsht\u00eb problematike, pasi nd\u00ebrsa motor\u00ebt e k\u00ebrkimit ofrojn\u00eb informacione q\u00eb mund t\u00eb verifikohen, nd\u00ebrsa chatbot-et p\u00ebrgjigjen me qasjen miq\u00ebsore t\u00eb nj\u00eb asistenti digjital, duke mashtruar potencialisht p\u00ebrdoruesit. Nj\u00eb mohim i shkurt\u00ebr \u00ebsht\u00eb i pamjaftuesh\u00ebm p\u00ebr t\u00eb kund\u00ebrshtuar k\u00ebt\u00eb problematik\u00eb.<\/p>\n\n\n\n
N\u00eb ndryshim nga ChatGPT, chatbot-i i Bing, i cili gjithashtu p\u00ebrdor t\u00eb nj\u00ebjt\u00ebn teknologji, performon m\u00eb mir\u00eb me detyrat e gjetjes s\u00eb fakteve, duke i lidhur ato me burimet. ChatGPT mund t\u00eb k\u00ebrkoj\u00eb n\u00eb m\u00ebnyr\u00eb t\u00eb ngjashme n\u00eb web, por kjo ve\u00e7ori disponohet vet\u00ebm n\u00eb versionin Plus, i cili \u00ebsht\u00eb me pages\u00eb. Rrjedhimisht, natyra e pavarur e ChatGPT e b\u00ebn at\u00eb m\u00eb t\u00eb prirur ndaj p\u00ebrdorues\u00ebve q\u00eb nuk mund t\u00eb verifikojn\u00eb gjithmon\u00eb informacionin.<\/p>\n\n\n\n
Nd\u00ebrsa OpenAI ka b\u00ebr\u00eb p\u00ebrmir\u00ebsime me ChatGPT q\u00eb nga publikimi i platform\u00ebs, ka ende mosp\u00ebrputhje t\u00eb konsiderueshme. P\u00ebr shembull, kur u pyet n\u00ebse mund t\u00eb zbuloj\u00eb tekstin e krijuar nga AI, roboti paralajm\u00ebroi se ai “nuk ishte i pagabuesh\u00ebm”, por m\u00eb pas pohoi gabimisht se teksti n\u00eb fjal\u00eb ishte krijuar nga AI. N\u00eb m\u00ebnyr\u00eb t\u00eb ngjashme, kur u pyet p\u00ebr rekomandime librash p\u00ebr ekonomin\u00eb, ai ofroi dy tituj krejt\u00ebsisht t\u00eb fabrikuar.<\/p>\n\n\n\n
Duke pasur parasysh k\u00ebto mosp\u00ebrputhje, kritik\u00ebt argumentojn\u00eb se mohimi i p\u00ebrgjegj\u00ebsis\u00eb, “Mund t\u00eb gjeneroj\u00eb her\u00eb pas here informacione t\u00eb pasakta”, duket se n\u00ebnvler\u00ebson shpesht\u00ebsin\u00eb e gabimeve. Si t\u00eb tilla, deklarata m\u00eb t\u00eb qarta nga OpenAI, t\u00eb shoq\u00ebruara me p\u00ebrmir\u00ebsime n\u00eb besueshm\u00ebrin\u00eb e sistemit, jan\u00eb t\u00eb nevojshme p\u00ebr t’u mbrojtur kund\u00ebr keqp\u00ebrdorimeve t\u00eb mundshme dhe pasojave t\u00eb d\u00ebmshme.<\/p>\n","protected":false},"excerpt":{"rendered":"
Chatbot-i i inteligjenc\u00ebs artificiale, ChatGPT, ka marr\u00eb v\u00ebmendje p\u00ebr prirjen e tij p\u00ebr t\u00eb ofruar informacione t\u00eb pasakta. Megjith\u00ebse OpenAI paralajm\u00ebron n\u00eb faqen kryesore se AI mund t\u00eb “gjeneroj\u00eb her\u00eb pas here informacione t\u00eb pasakta”, shum\u00eb kritik\u00eb argumentojn\u00eb se ky paralajm\u00ebrim \u00ebsht\u00eb i pamjaftuesh\u00ebm, duke pasur parasysh pasojat e mundshme t\u00eb keqinformimit. Kufizimet e chatbot-it<\/p>\n","protected":false},"author":4,"featured_media":13424,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[38],"tags":[40],"class_list":{"0":"post-13595","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-tech","8":"tag-featured"},"yoast_head":"\n