Kwixesha elidlulileyo khange sithethe ngayo I-OpenAI kunye nerhanisi lakhe elizalela amaqanda egolide, ChatGPT. Kwaye kukuba iindaba zamva nje azinanto yakwenza nokusebenza kwazo kwaye kuninzi okunokwenza nokuba nokwenzeka kokuthintela i-chatbot edumileyo. Isizathu, ukuba ubumfihlo abunakuqinisekiswa, ukuba inkampani inokwazi yonke into ngathi kuba sibaxelela nantoni na. Kule veki, inkampani ibonakala ithathe inyathelo lokuthintela iingxaki namanye amazwe.
Le yindlela i-OpenAI eyenza ngayo ukuba yaziwe kwinqaku iposwe ngoAprili 25, ngaphantsi kweeyure ezingama-24 ezidlulileyo. Isihloko sayo sithi "Iindlela ezintsha zokulawula idatha yakho kwi-ChatGPT", kwaye umxholo ngokusisiseko yinkcazo yokhetho olutsha apho singenza imbali yengxoxo ingasebenzi. Singakwazi ukufikelela kuyo ngokukhetha umsebenzisi wethu (ezantsi ngasekhohlo), ucofa "Useto", kwi "Data controls" ucofa "Bonisa" kwaye uvale iswitshi.
I-OpenAI inokuba neengxaki ukuba ayiqinisekisi ubumfihlo
Isicatshulwa esicacisayo phantsi kweso sitshixo sithi [ukuba kukhangelwe) iingxoxo ezintsha zigcinwa kwimbali yethu kwaye siyazivumela ukuba zisetyenziswe ukuqeqesha iChatGPT. Iincoko ezingagcinwanga ziya kususwa kwiinkqubo ze-OpenAI kwiintsuku ezingama-30. Nangona kunjalo, iingxoxo ziya kugcinwa ngaloo nyanga, kodwa, bathi, ukuphonononga kuphela xa kuyimfuneko ukuqinisekisa ukuba akukho kuxhatshazwa.
Olunye ukhuseleko lwabasebenzisi kukuba, xa sisebenzisa i-ChatGPT ngenkqubo okanye ngenkonzo esebenzisa eyakho Business APIs, idatha yomsebenzisi ayizukusetyenziswa ukuqeqesha i-chatbot ngokungagqibekanga. Oku kuya kuza kwiinyanga ezimbalwa ezizayo.
Into yokugqibela abasixelela ngayo kweli nqaku ukuba kunokwenzeka ukuthumela ngaphandle idatha yethu yokusetyenziswa kwi-ChatGPT kwaye uqonde ukuba loluphi ulwazi olugcina i-ChatGPT. Ukuba sithatha isigqibo sokuzithumela ngaphandle, siya kufumana ulwazi kwi-imeyile yethu yokubhalisa. Ixesha lokufumana loo imeyile liya kuxhomekeka kwizinto ezingachazwanga, kodwa ndiyicelile kwaye, emva kwemizuzu emininzi yokulinda, andikafumani nto.
Kusenokuba ngenxa yezinye izizathu, kodwa kubonakala kum ukuba le ntshukumo iyimpendulo kwizisongelo zokuvala i-blockade eziwiswe ngamazwe afana ne-Italy, iJamani kunye neSpain, eyayijonge ukubuza ukuba yintoni eyenziwa ngabamelwane bethu baseYurophu.
Icebiso elinye: ulawulo lwembali yesandla
Ndingacebisa ngenye indlela. Njengomsebenzisi obesebenzisa i-OpenAI chatbot ukubuza imibuzo emininzi, nokuba athethe malunga neengoma zeengoma okanye iintsingiselo zemuvi, imbali yam yayizaliswe ziincoko ezingenakubalwa. Inyani kukuba unokumbuza "Yintoni igama lomculi we-Ad Infinitum" kwaye oko kugcinwa njengengxoxo, nokuba ngumbuzo kunye nempendulo yokugcina ixesha. Ndisanda kunika cima yonke imbali kwaye ndiyiqalile ukusuka ekuqaleni.
Olunye ukhetho asikokugcina nantoni na, kodwa ukuba sithathe into esinomdla kuyo, asinakuphinda siyijonge kwakhona. Ke ngoko, umbono olungileyo, okanye kubonakala kum, kukwenza kanye le nto ndiyenzayo ngoku: Ndiqala iincoko eziqhelekileyo, kwaye xa ukuthandabuza kwam kucinyiwe okanye ndifumana into endiyifunayo, Ndiyicima ngesandla. Kwaye andiyenzeli nje ubumfihlo, kodwa ukuba yonke into icwangciswe ngakumbi. Ewe, ukuba ndiziva ngathi ndibuza into engaqhelekanga okanye eyezonyango (nangona kungcono ukuya kwingcali, yinto esele siyenzile ngoGoogle), ndiyayicima loo ncoko ukuba ibe yimfihlo.
Nokuba yeyiphi na into esetyenziswayo, le nto inokwenzeka sele ikhona, kwaye ngoku kuxhomekeke kumazwe ukuba aqinisekise ukuba oko bakuthethayo kuyinyani kwaye kwanele.