Studiuesit e siguris\u00eb dhe teknolog\u00ebt q\u00eb hetojn\u00eb chatbot-et e personalizuara i kan\u00eb b\u00ebr\u00eb ata t\u00eb zbulojn\u00eb udh\u00ebzimet fillestare q\u00eb u jan\u00eb dh\u00ebn\u00eb kur u krijuan, dhe gjithashtu kan\u00eb zbuluar dhe shkarkuar skedar\u00ebt e p\u00ebrdorur p\u00ebr t\u00eb personalizuar chatbot-et. Informacioni personal i njer\u00ebzve ose t\u00eb dh\u00ebnat pron\u00ebsore mund t\u00eb vihen n\u00eb rrezik, thon\u00eb ekspert\u00ebt.<\/p>\n\n\n\n
“Shqet\u00ebsimet e privat\u00ebsis\u00eb s\u00eb rrjedhjes s\u00eb skedar\u00ebve duhet t\u00eb merren seriozisht,” thot\u00eb Jiahao Yu, nj\u00eb studiues i shkencave kompjuterike n\u00eb Universitetin Northwestern. “Edhe n\u00ebse nuk p\u00ebrmbajn\u00eb informacione t\u00eb ndjeshme, ato mund t\u00eb p\u00ebrmbajn\u00eb disa njohuri q\u00eb projektuesi nuk d\u00ebshiron t’i ndaj\u00eb me t\u00eb tjer\u00ebt dhe [q\u00eb sh\u00ebrben] si pjes\u00eb thelb\u00ebsore e GPT-s\u00eb me porosi.”<\/p>\n\n\n\n
S\u00eb bashku me studiues t\u00eb tjer\u00eb n\u00eb Northwestern, Yu ka testuar m\u00eb shum\u00eb se 200 GPT t\u00eb personalizuara dhe e ka gjetur “\u00e7udit\u00ebrisht t\u00eb drejtp\u00ebrdrejt\u00eb” zbulimin e informacionit prej tyre. “Shkalla jon\u00eb e suksesit ishte 100 p\u00ebr qind p\u00ebr rrjedhjen e skedar\u00ebve dhe 97 p\u00ebr qind p\u00ebr nxjerrjen e menj\u00ebhershme t\u00eb sistemit, e arritshme me k\u00ebrkesa t\u00eb thjeshta q\u00eb nuk k\u00ebrkojn\u00eb njohuri t\u00eb specializuara n\u00eb inxhinierin\u00eb e shpejt\u00eb ose grupimin e shpejt\u00eb,” thot\u00eb Yu.<\/p>\n\n\n\n
GPT-t\u00eb e personalizuara, p\u00ebr nga dizajni i tyre, jan\u00eb t\u00eb lehta p\u00ebr t’u b\u00ebr\u00eb. Njer\u00ebzit me nj\u00eb abonim OpenAI jan\u00eb n\u00eb gjendje t\u00eb krijojn\u00eb GPT, t\u00eb cilat njihen gjithashtu si agjent\u00eb t\u00eb AI. OpenAI thot\u00eb se GPT-t\u00eb mund t\u00eb nd\u00ebrtohen p\u00ebr p\u00ebrdorim personal ose t\u00eb publikohen n\u00eb ueb. Kompania planifikon q\u00eb zhvilluesit p\u00ebrfundimisht t\u00eb jen\u00eb n\u00eb gjendje t\u00eb fitojn\u00eb para n\u00eb var\u00ebsi t\u00eb numrit t\u00eb njer\u00ebzve q\u00eb p\u00ebrdorin GPT-t\u00eb.<\/p>\n","protected":false},"excerpt":{"rendered":"
Studiuesit e siguris\u00eb dhe teknolog\u00ebt q\u00eb hetojn\u00eb chatbot-et e personalizuara i kan\u00eb b\u00ebr\u00eb ata t\u00eb zbulojn\u00eb udh\u00ebzimet fillestare q\u00eb u jan\u00eb dh\u00ebn\u00eb kur u krijuan, dhe gjithashtu kan\u00eb zbuluar dhe shkarkuar skedar\u00ebt e p\u00ebrdorur p\u00ebr t\u00eb personalizuar chatbot-et. Informacioni personal i njer\u00ebzve ose t\u00eb dh\u00ebnat pron\u00ebsore mund t\u00eb vihen n\u00eb rrezik, thon\u00eb ekspert\u00ebt. “Shqet\u00ebsimet e<\/p>\n","protected":false},"author":4,"featured_media":16286,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[349],"tags":[43],"class_list":{"0":"post-16284","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-artificial-intelligence","8":"tag-trending"},"yoast_head":"\n