Kusukela ku《Odyssey》 kuya kumanoveli obugebengu ka-Agatha Christie, abantu badale izindaba eziningi ezimangalisayo ezigcwalise amashalofu ezizukulwane eziningi. Kodwa ucwaningo lubonisa ukuthi manje ubuhlakani bokwenziwa kungenzeka buqale ukuba yimbangi yababhali abangabantu.
Ucwaningo olushicilelwe kumagazini i-Judgment and Decision Making lubonise ukuthi abacwaningi bameme abantu abadala abangu-1,682 ukuba bafunde izindaba ezimfishane ezingu-6, lapho ezingu-3 zazibhalwe abantu kanti ezinye ezingu-3 zazibhalwe yi-ChatGPT. Indaba ngayinye ye-AI yayigxile endikimbeni efana naleyo kwenye yezindaba ezibhalwe umuntu.
Abacwaningi batshela ababambiqhaza ukuthi izindaba abazifundayo zibhalwe umuntu noma yi-AI, kodwa lolu lwazi lwalungahlali lunembile. Ngemva kwalokho, bacela ababambiqhaza ukuba bahlole ukuheha kwendaba, ulwazi lokufunda kanye nekhwalithi iyonke.
Imiphumela yabonisa ukuthi ababambiqhaza abafunda izindaba ezikhiqizwe yi-AI babecabanga ukuthi lezi zindaba ziheha kakhulu futhi zisezingeni eliphezulu kunalabo abafunda izindaba ezibhalwe abantu. Nokho, lapho ababambiqhaza betshelwa ukuthi indaba ethile ibhalwe umuntu, babeyinika amaphuzu aphezulu kakhulu.
Umbhali oyinhloko walolu cwaningo, uDkt. Deena Skolnick Weisberg wase-Villanova University ePennsylvania, e-United States, uthe: “Izinhlelo ze-AI sezikwazi ukubhala izindaba ezimfishane, futhi abantu bacabanga ukuthi lezi zindaba okungenani azikho ngaphansi kwemisebenzi ebhalwe abantu, kanti kungenzeka zibe ngcono. Kufanele sisebenzise lokhu ukuze siqonde kabusha amakhono e-AI.”
Nokho, uWeisberg naye ungumdali. Uthe lokhu akusho ukuthi ukubhala kufanele kudluliselwe ku-AI. Uveze ukuthi amanoveli akhiqizwa izinkampani zobuchwepheshe zisebenzisa i-AI kungenzeka ahluke emisebenzini yababhali abangabantu.
Uthe: “Kungenzeka sidinge ukunika indawo amanoveli abhalwe yi-AI, kanye namanoveli abhalwe ngokubambisana yi-AI nabantu, kodwa lokhu akusho ukuthi inqubo yokudala yabantu ayisafanele ukwaziswa. Abantu babhala ukuze baziveze ngobuciko, baziphonsele inselelo futhi baqonde futhi bahlele amava abo. Iqiniso lokuthi i-AI ingakhiqiza izindaba ezifana nalezo ezenziwa abantu aliziguquli lezi zinto.”
Imiphumela yocwaningo iphinde yabonisa ukuthi isimo sengqondo sabantu nge-AI siyawathinta amaphuzu abawanikezayo. Ababambiqhaza ababenesimo sengqondo esihle kakhulu nge-AI banikeze amaphuzu aphezulu kunalabo ababenesimo sengqondo esingesihle kakhulu nge-AI lapho betshelwa ukuthi indaba ibhalwe yi-ChatGPT.
Abacwaningi babe sebeqhuba ezinye izivivinyo ezimbili, lapho kwaba nabantu abadala abangu-905. Ababambiqhaza babebona ezimbili kwezingu-6 zezindaba ezimfishane. Lezi zindaba ezimbili zazinendikimba efanayo, kodwa eyodwa kuphela yazo yayibhalwe umuntu. Ababambiqhaza kwakudingeka banqume ukuthi iyiphi ebhalwe umuntu nokuthi iyiphi ekhiqizwe yi-AI.
Kwesinye salezi zivivinyo, cishe u-40% wababambiqhaza waphendula kahle, okuyizinga eliphansi ngisho kunokuqagela ngokungahleliwe; kwesinye isivivinyo izinga lokuphendula kahle laba ngu-52%.
Abacwaningi babhala: “Abukho ubufakazi obujwayelekile obubonisa ukuthi ababambiqhaza bangakwazi ukuhlukanisa phakathi kwezindaba ezibhalwe abantu nezikhiqizwe yi-AI.”
Abacwaningi baphinde bathola ukuthi ulwazi lokufunda noma lokubhala amanoveli ababambiqhaza ababezibika lona lwalungahlobene ngokusobala nokunemba kwabo ekwahluleleni ukuthi ubani umbhali wendaba; kodwa abantu ababejwayelene kakhulu nezinhlelo ze-AI babenokunemba okuphezulu.
UWeisberg uthe: “Ukubhala kwe-AI kunesitayela esithile, futhi ngokuzijwayeza singafunda kancane kancane ukusibona, njengoba singafunda ukubona izitayela zokubhala ezihlukile zababhali abangabantu.”
UWeisberg ukholelwa ukuthi ukuthandwa kakhulu kwezindaba ezibhalwe ukuthi “zibhalwe umuntu” kungenzeka kuhlobene nokufuna kwabantu ubuqiniso; kodwa ngokolwazi lokufunda uqobo, isizathu esenza abantu bathande izindaba ezikhiqizwe yi-AI kakhulu kungenzeka ukuthi kulula ukuzifunda nokuziqonda.
Nokho, uLuke Kennard, uprofesa wefilimu nokubhala ngobuciko e-University of Birmingham e-United Kingdom, obengabambi qhaza kulolu cwaningo, ugcizelele ukuthi i-AI iseyisihloko esinempikiswano enkulu.
Uthe: “Ngenxa yokuthi singabantu, sivame ukwenza i-AI ibonakale njengomuntu, siyicabange njengerobhothi elikhanyayo elinobuso obunomusa nobungathathi hlangothi. Kodwa empeleni, iyikhodi ebikezelayo esebenza kusizindalwazi esikhulu semibhalo eyebiwe, futhi le khodi isebenza ezikhungweni zedatha ezinkulu kakhulu, ezibiza imali eningi futhi ezilimazayo, ezidala imiphumela eyinhlekelele emiphakathini eseduze.”
UKennard uphinde wagcizelela ukuthi i-AI ayilona “ithuluzi” elijwayelekile elifana nethuluzi lokuhlola isipelingi noma isichazamazwi samagama afanayo, kodwa iwusongo oluthinta ukusinda.
Uthe: “Yebo, uma umyalo ubhalwe kahle futhi kwenziwa ukulungisa nokupholisha okuthile, ingakhiqiza inkondlo ehambisanayo, mhlawumbe ejabulisayo. Phela, idatha yayo yokuqeqesha ivela emiphumeleni yokudala yabantu yezinkulungwane zeminyaka; uma ingakwazi ngisho nalokho, lokho bekuyoba yihlazo. Ngicabanga ukuthi umbuzo okufanele siwubuze ngempela uthi: Ingabe konke lokhu kuyafaneleka? Impendulo ngokuqinisekile ithi, ngaphandle kokungabaza, cha — akufanele.”