Kompanit\u00eb q\u00eb merren me inteligjenc\u00ebn artificiale po p\u00ebrballen me nj\u00eb sfid\u00eb me t\u00eb cil\u00ebn njer\u00ebzit jan\u00eb p\u00ebrballur q\u00eb nga ekzistenca e tyre: si t\u00eb ruajn\u00eb kujtimet.<\/p>\n
OpenAI, Google, Meta dhe Microsoft jan\u00eb p\u00ebrqendruar muajt e fundit n\u00eb memorien, duke njoftuar p\u00ebrdit\u00ebsime q\u00eb u mund\u00ebsojn\u00eb chatbot-\u00ebve t\u00eb ruajn\u00eb m\u00eb shum\u00eb informacione p\u00ebr p\u00ebrdoruesit, n\u00eb m\u00ebnyr\u00eb q\u00eb t\u2019u japin p\u00ebrgjigje m\u00eb t\u00eb personalizuara.<\/p>\n
Ky hap konsiderohet si nj\u00eb zhvillim i r\u00ebnd\u00ebsish\u00ebm p\u00ebr t\u00eb ndihmuar grupet kryesore n\u00eb fush\u00ebn e inteligjenc\u00ebs artificiale t\u00eb t\u00ebrheqin p\u00ebrdorues n\u00eb nj\u00eb treg konkurrues t\u00eb chatbot-\u00ebve dhe agjent\u00ebve, si dhe si nj\u00eb mjet p\u00ebr t\u00eb gjeneruar t\u00eb ardhura nga teknologjit\u00eb m\u00eb t\u00eb avancuara
\nMegjithat\u00eb, kritik\u00ebt kan\u00eb paralajm\u00ebruar se ky zhvillim mund t\u00eb keqp\u00ebrdoret p\u00ebr t\u00eb shfryt\u00ebzuar p\u00ebrdoruesit p\u00ebr q\u00ebllime komerciale dhe se mund t\u00eb shkaktoj\u00eb shqet\u00ebsime lidhur me privat\u00ebsin\u00eb, shkruan Financial Times.<\/p>\n
Identifikimi i kontekstit nga t\u00eb dh\u00ebnat e jashtme<\/p>\n
\u201cN\u00ebse ke nj\u00eb agjent q\u00eb t\u00eb njeh v\u00ebrtet, sepse ka mbajtur mend bisedat, e gjith\u00eb sh\u00ebrbimi b\u00ebhet m\u00eb t\u00ebrheq\u00ebs, dhe kur fillon t\u00eb p\u00ebrdor\u00ebsh nj\u00eb produkt t\u00eb till\u00eb, nuk kalon m\u00eb te nj\u00eb tjet\u00ebr,\u201d tha Pattie Maes, profesoresh\u00eb n\u00eb Laboratorin e Medias n\u00eb MIT dhe eksperte p\u00ebr nd\u00ebrveprimin mes njer\u00ebzve dhe inteligjenc\u00ebs artificiale.<\/p>\n
Chatbot-\u00ebt me inteligjenc\u00eb artificiale si Gemini i Google-it dhe ChatGPT i OpenAI-t kan\u00eb sh\u00ebnuar p\u00ebrparime t\u00eb m\u00ebdha, transmeton Telegrafi.<\/p>\n
P\u00ebrmir\u00ebsimet p\u00ebrfshijn\u00eb zgjerimin e dritares s\u00eb kontekstit, q\u00eb p\u00ebrcakton sa shum\u00eb biseda mund t\u00eb mbaj\u00eb mend nj\u00eb chatbot nj\u00ebkoh\u00ebsisht, si dhe p\u00ebrdorimin e teknikave si gjenerimi p\u00ebr t\u00eb identifikuar kontekstin p\u00ebrkat\u00ebs nga t\u00eb dh\u00ebnat e jashtme.<\/p>\n
Organizatat q\u00eb merren me inteligjenc\u00ebn artificiale gjithashtu kan\u00eb p\u00ebrmir\u00ebsuar memorien afatgjat\u00eb t\u00eb modeleve AI duke ruajtur profilet dhe preferencat e p\u00ebrdoruesve, p\u00ebr t\u00eb ofruar p\u00ebrgjigje m\u00eb t\u00eb dobishme dhe m\u00eb t\u00eb personalizuara.<\/p>\n
P\u00ebr shembull, nj\u00eb chatbot mund t\u00eb mbaj\u00eb mend n\u00ebse p\u00ebrdoruesi \u00ebsht\u00eb vegjetarian dhe t\u00eb p\u00ebrgjigjet n\u00eb p\u00ebrputhje me k\u00ebt\u00eb kur jep rekomandime p\u00ebr restorante ose receta.<\/p>\n
N\u00eb muajin mars, Google zgjeroi memorien e Geminit duke p\u00ebrfshir\u00eb historikun e k\u00ebrkimeve t\u00eb p\u00ebrdoruesit, me kusht q\u00eb personi t\u00eb pranoj\u00eb k\u00ebt\u00eb, ndryshe nga kufizimi i m\u00ebparsh\u00ebm q\u00eb ishte vet\u00ebm bisedat me chatbotin, dhe planifikon ta zgjas\u00eb k\u00ebt\u00eb n\u00eb aplikacionet e tjera t\u00eb Google n\u00eb t\u00eb ardhmen.<\/p>\n
\u201cSa m\u00eb mir\u00eb q\u00eb chatboti ju njeh, aq m\u00eb mir\u00eb do t\u00eb ofroj\u00eb sh\u00ebrbimin\u201d<\/p>\n
\u201cAshtu si me nj\u00eb asistent njer\u00ebzor, sa m\u00eb shum\u00eb t\u00eb kuptoj\u00eb q\u00ebllimet tuaja, kush jeni dhe \u00e7far\u00eb \u00ebsht\u00eb e r\u00ebnd\u00ebsishme p\u00ebr ju, aq m\u00eb t\u00eb mir\u00eb sh\u00ebrbim do t\u00eb mund t\u00eb ofroj\u00eb,\u201d tha Michael Siliski, drejtor i menaxhimit t\u00eb produkteve n\u00eb Google DeepMind.<\/p>\n
ChatGPT i OpenAI-t dhe chatbot-i i Met\u00ebs n\u00eb WhatsApp dhe Messenger mund t\u00eb referohen n\u00eb bisedat e kaluara, jo vet\u00ebm n\u00eb sesionin aktual. P\u00ebrdoruesit mund t\u00eb fshijn\u00eb kujtime t\u00eb caktuara n\u00eb cil\u00ebsimet dhe informohen kur modeli krijon nj\u00eb kujtim me nj\u00eb mesazh n\u00eb ekran.<\/p>\n
\u201cMemoria ndihmon ChatGPT-n t\u00eb b\u00ebhet m\u00eb i dobish\u00ebm me kalimin e koh\u00ebs duke i b\u00ebr\u00eb p\u00ebrgjigjet m\u00eb relevante\u201d, njoftoi kompania OpenAI dhe shtoi se \u201cju gjithmon\u00eb keni kontroll dhe mund t\u2019i pyesni ChatGPT-n se \u00e7far\u00eb kujton p\u00ebr ju, t\u00eb ndryshoni kujtimet e ruajtura dhe bisedat e kaluara, ose t\u00eb \u00e7aktivizoni memorien n\u00eb \u00e7do moment\u201d.<\/p>\n
P\u00ebr p\u00ebrdoruesit t\u00eb bizneseve, Microsoft p\u00ebrdor t\u00eb dh\u00ebna organizative p\u00ebr t\u00eb informuar memorien, si email-e, kalendar\u00eb dhe dokumente t\u00eb brendshme. Gjiganti teknologjik n\u00eb prill nisi t\u00eb ofroj\u00eb n\u00eb disa pajisje funksionin Preview t\u00eb ve\u00e7oris\u00eb Recall, q\u00eb regjistron aktivitetet e p\u00ebrdoruesve duke marr\u00eb pamje nga ekrani i kompjuterit. P\u00ebrdoruesit mund ta \u00e7aktivizojn\u00eb k\u00ebt\u00eb funksion ose t\u00eb ndalojn\u00eb p\u00ebrkoh\u00ebsisht regjistrimin e ekranit.<\/p>\n
Rritja e monetizimit p\u00ebrmes marketingut partner<\/p>\n
Kur kjo ve\u00e7ori u njoftua p\u00ebr her\u00eb t\u00eb par\u00eb n\u00eb maj t\u00eb vitit t\u00eb kaluar, shkaktoi shqet\u00ebsime n\u00eb komunitetin e siguris\u00eb kibernetike dhe tek t\u00eb tjer\u00eb q\u00eb e konsideruan t\u00eb frikshme, gj\u00eb q\u00eb \u00e7oi n\u00eb disa vonesa t\u00eb nisjes s\u00eb saj.<\/p>\n
Kompanit\u00eb e AI gjithashtu besojn\u00eb se memoria e p\u00ebrmir\u00ebsuar mund t\u00eb luaj\u00eb nj\u00eb rol t\u00eb madh n\u00eb rritjen e monetizimit p\u00ebrmes marketingut partner dhe reklamave.<\/p>\n
Drejtori ekzekutiv i Met\u00ebs, Mark Zuckerberg, s\u00eb fundmi tha se do t\u00eb ket\u00eb nj\u00eb mund\u00ebsi t\u00eb madhe p\u00ebr shfaqjen e rekomandimeve t\u00eb produkteve ose reklamave n\u00eb chatbot-in e tyre.<\/p>\n
N\u00eb muajin prill, OpenAI p\u00ebrmir\u00ebsoi funksionalitetet e tij p\u00ebr blerje n\u00eb ChatGPT p\u00ebr t\u00eb shfaqur m\u00eb mir\u00eb produkte, sh\u00ebrbime dhe recensione. Kompania i tha Financial Times se aktualisht nuk ka lidhje partneriteti.<\/p>\n
Megjithat\u00eb, futja e kapacitetit m\u00eb t\u00eb madh t\u00eb memories n\u00eb modelet e m\u00ebdha gjuh\u00ebsore ka shkaktuar gjithashtu shqet\u00ebsime p\u00ebr privat\u00ebsin\u00eb, pasi rregullator\u00ebt n\u00eb mbar\u00eb bot\u00ebn po i v\u00ebzhgojn\u00eb me kujdes m\u00ebnyrat se si modelet mund t\u00eb manipulojn\u00eb p\u00ebrdoruesit p\u00ebr fitim.<\/p>\n
Rreziku nga paragjykimet ose gabimet<\/p>\n
Memoria e zgjeruar gjithashtu mund t\u00eb \u00e7oj\u00eb n\u00eb at\u00eb q\u00eb modelet p\u00ebrpiqen tep\u00ebr t\u00eb p\u00ebrshtatin p\u00ebrgjigjet e tyre me preferencat e p\u00ebrdoruesit, duke forcuar k\u00ebshtu paragjykimet ose gabimet.<\/p>\n
OpenAI k\u00ebrkoi falje pasi u zbulua se modeli i tij GPT-4o ishte tep\u00ebr i komplimentuesh\u00ebm dhe i ndikuar leht\u00eb, dhe modeli u kthye n\u00eb versionin e m\u00ebparsh\u00ebm.<\/p>\n
N\u00eb m\u00ebnyr\u00eb m\u00eb t\u00eb p\u00ebrgjithshme, modelet e inteligjenc\u00ebs artificiale mund t\u00eb \u201chalucionojn\u00eb\u201d, duke krijuar p\u00ebrgjigje t\u00eb pav\u00ebrteta ose t\u00eb pakuptimta, dhe p\u00ebrjetojn\u00eb nj\u00eb \u201czhvendosje memorjeje\u201d kur kujtimet b\u00ebhen t\u00eb vjetruara ose kontradiktore, q\u00eb ndikon n\u00eb sakt\u00ebsin\u00eb.<\/p>\n
\u201cSa m\u00eb shum\u00eb q\u00eb sistemi di p\u00ebr ju, aq m\u00eb shum\u00eb mund t\u00eb p\u00ebrdoret n\u00eb q\u00ebllime negative \u2014 p\u00ebr t\u2019ju b\u00ebr\u00eb t\u00eb blini di\u00e7ka ose p\u00ebr t\u2019ju bindur p\u00ebr besime t\u00eb caktuara. Prandaj duhet t\u00eb filloni t\u00eb mendoni p\u00ebr interesat themelore t\u00eb kompanive q\u00eb ofrojn\u00eb k\u00ebto sh\u00ebrbime,\u201d tha Maes, profesoresh\u00eb n\u00eb MIT.<\/p>\n
The post Pse kujtesa \u00ebsht\u00eb b\u00ebr\u00eb \u00e7el\u00ebsi kryesor i \u2018chatbot\u2019 t\u00eb inteligjenc\u00ebs artificiale?<\/a> appeared first on www.lexo.al<\/a>.<\/p>","protected":false},"excerpt":{"rendered":" Kompanit\u00eb q\u00eb merren me inteligjenc\u00ebn artificiale po p\u00ebrballen me nj\u00eb sfid\u00eb me t\u00eb cil\u00ebn njer\u00ebzit jan\u00eb p\u00ebrballur q\u00eb nga ekzistenca e tyre: si t\u00eb ruajn\u00eb kujtimet. OpenAI, Google, Meta dhe Microsoft jan\u00eb p\u00ebrqendruar muajt e fundit n\u00eb memorien, duke njoftuar p\u00ebrdit\u00ebsime q\u00eb u mund\u00ebsojn\u00eb chatbot-\u00ebve t\u00eb ruajn\u00eb m\u00eb shum\u00eb informacione p\u00ebr p\u00ebrdoruesit, n\u00eb m\u00ebnyr\u00eb q\u00eb […]<\/p>\n","protected":false},"author":0,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-22770","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/zennihost.com\/test\/wp-json\/wp\/v2\/posts\/22770","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/zennihost.com\/test\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/zennihost.com\/test\/wp-json\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/zennihost.com\/test\/wp-json\/wp\/v2\/comments?post=22770"}],"version-history":[{"count":0,"href":"https:\/\/zennihost.com\/test\/wp-json\/wp\/v2\/posts\/22770\/revisions"}],"wp:attachment":[{"href":"https:\/\/zennihost.com\/test\/wp-json\/wp\/v2\/media?parent=22770"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/zennihost.com\/test\/wp-json\/wp\/v2\/categories?post=22770"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/zennihost.com\/test\/wp-json\/wp\/v2\/tags?post=22770"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}