Iziphumo zeChatGPT

Kwi-cybersecurity

Hlola iPoint Research kunye nabanye baqaphele ukuba i-ChatGPT ikwazi ukubhalaphishingemails kunyei-malware, ngakumbi xa idibene neOpenAI Codex.I-CEO ye-OpenAI ibhale ukuba ukuqhubela phambili kwesoftware kunokubeka "(umzekelo) umngcipheko omkhulu wokhuseleko lwe-cyber" kwaye uqhubeke nokuqikelela "sinokufikelela kwi-AGI yokwenyani (ubukrelekrele jikelele bokwenziwa) kule minyaka ilishumi izayo, ngoko ke kufuneka sithathe umngcipheko woko kubaluleke kakhulu”.U-Altman waxoxa ukuba, ngelixa i-ChatGPT "ngokucacileyo ayisondelanga kwi-AGI", umntu kufuneka "ayithembei-exponential.Ujonge ngasemva,nkqo ujonge phambili.”

Kwimfundo ephakamileyo

I-ChatGPT inokubhala intshayelelo kunye namacandelo angabonakaliyo kumanqaku enzululwazi, aphakamisa imibuzo yokuziphatha.Amaphepha aliqela sele edwelisile iChatGPT njengombhali osebenzisanayo.

KwiIAtlantikiimagazini,UStephen Marcheyaqaphela ukuba impembelelo yayo kwizifundiswa kwaye ngakumbiizincoko isiceloiseza kuqondwa.Utitshala wesikolo esiphakamileyo saseCalifornia kunye nombhali uDaniel Herman wabhala ukuba i-ChatGPT iya kuqalisa "ukuphela kwesiNgesi samabanga aphakamileyo".KwiIndalokwijenali, uChris Stokel-Walker wachaza ukuba ootitshala kufuneka bazixhalabise malunga nabafundi abasebenzisa i-ChatGPT ukuze babhale ngaphandle, kodwa ababoneleli bezemfundo baya kuziqhelanisa nokuphucula ukucinga okuphangaleleyo okanye ukuqiqa.Emma Bowman kunyei-NPRwabhala ngengozi yabafundi abakopa ngesixhobo se-AI esinokuthi sikhuphe isicatshulwa esinomkhethe okanye esingenangqiqo ngethoni egunyazisiweyo: "Kusekho iimeko ezininzi apho ubuza umbuzo kwaye uya kukunika impendulo evakalayo engalunganga."

UJoanna Stern kunyeIWall Street Journalwachaza ukukopela kwisikolo samabanga aphakamileyo saseMelika ngesixhobo ngokungenisa isincoko esiveliswayo.Unjingalwazi uDarren Hick weIYunivesithi yaseFurmanichaze ngokuqaphela “isimbo” se-ChatGPT kwiphepha elingeniswe ngumfundi.I-detector ye-GPT ye-intanethi yabanga ukuba iphepha laliyi-99.9 yeepesenti ukuba lenziwe ngekhompyutha, kodwa u-Hick wayengenabo ubungqina obuqinileyo.Nangona kunjalo, umfundi ekuthethwa ngaye wavuma ukusebenzisa i-GPT xa ejongene naye, kwaye ngenxa yoko wasilela ikhosi.U-Hick ucebise umgaqo-nkqubo wokunika i-ad-hoc yovavanyo lomlomo lomntu ngamnye kwisihloko sephepha ukuba umfundi ukrokrelwa ngamandla ukuba angenise iphepha elenziwe nge-AI.U-Edward Tian, ​​umfundi onesidanga esiphezulu eIYunivesithi yasePrinceton, idale inkqubo, ebizwa ngokuba yi "GPTZero," emisela ukuba ingakanani isicatshulwa esenziwe yi-AI, iboleka ukuba isetyenziswe ukukhangela ukuba isincoko sibhalwe ngumntu ukulwa.ukukopa ngokwemfundo.

Ukusukela nge-4 kaJanuwari 2023, iSebe lezeMfundo lesiXeko saseNew York likuthintele ukufikelela kwi-ChatGPT ukusuka kwi-intanethi yesikolo sikarhulumente kunye nezixhobo.

Kuvavanyo olumfamekileyo, i-ChatGPT yagwetywa ukuba iphumelele iimviwo zenqanaba lesidangaIYunivesithi yaseMinnesotakwinqanaba lomfundi we-C + nakwiIsikolo saseWharton kwiYunivesithi yasePennsylvaniaukusuka ku-B ukuya kwibakala B.(iWikipedia)

Ngexesha elizayo siza kuthetha malunga neenkxalabo zokuziphatha ze-ChatGPT.