Abavoti kukhetho lukamongameli weMelika basengozini yokuphiwa imibiko yamanga ngabasebenzisa imitshina yakulezinsuku i-Artificial Intelligence (AI) ukulingisela abantu ngenhloso yokuthuthukisa abanye abezombusazwe kumbe ohulumende bamanye amazwe.
Ukusetshenziswa kwe-Artificial intelligence sekuqale ukubonakala kulolukhetho kulandela ukuphuma komfanekiso owalungiswa nge-AI okaDonald Trump esanga ingcwethi yezemikhuhlane ethelelanwayo uDokotela Anthony Fauci kanye lemifanekiso yamanga eveza uJoe Biden etshela abavoti beNew Hampshire ukuthi batshiyane lokhetho lwamaprimary.
Ingcwethi yezokukhutshwa kwemibiko uElaine Kamarck oweBrookings Institution uthi isikhathi esiphuma imifanekiso lamavidiyo amanga la yiso esenza abeyingozi kakhulu.
“Umuntu uyenelisa ukukhupha imibiko yamanga sekusele amahola angamatshumi amabili lane ukuthi kuqhutshwe ukhetho. Kunzima ukulwisana lemibiko enje, njalo ilakho ukuletha umahluko omkhulu kukhetho.”
Inkampani kaRijul Gupta eyeDeepMedia ephathisa isigodlo somvikela esePentagon ukubamba imifanekiso, amazwi lamavidiyo amanga ithi avamile ngoba engadingi imali enengi ukulungisa njalo elula ukulungisa.
“Kuthatha imizuzu elitshumi lanhlanu kuphela ukulungisa imifanekiso le. Kulendlela zokulungisa imifanekiso le kungela mbadalo ezitholakala ebulenjini njalo nxa umuntu efuna ukubhadala ulakho ukubhadala amacent amabili kuphela ukulungisa ilizwi elilemizuzu engamatshumi amathathu.”
Ingqwetha elwela inkululeko yabakhupha imibiko uAri Cohn uthi.
"Kukhulunywa kakhulu ngeRussia kodwa iChina layo ilwa isebenzisa i-AI izama ukuthi ikhokhele abanye."
UCohn uthi inhlelo eziqhutshwa ngaphandle kwelizwe zihlose ukupha izizalwane zeMelika ezilodumo imibiko yamanga ukuze zibe yizo eziyithumela abanye.
Amankampani amakhulu awemitshina yakulezinsuku kanye lawebulenjini ethembisa ukulwisa ukusetshenziswa kwe-AI kumikhankaso yekhetho, kodwa akukabi lomthetho welizwe okhangela loludaba.
“Kunzima ukubeka umthetho okhangela imibiko yezombusazwe eMelika. I-First Amendment ivikela imibiko yezombusazwe, njalo lokhu kulesizatho esihle ngoba asifuni ukuthi uhulumende angenele kukhetho.”
UGupta uthi udubo olukhulu ngolokwahlukanisa umbiko wamanga loweqiniso. Pho singabakwazi njani nxa umbiko ungowamanga i-deepfake?
"Amazwi aphuma ehayiza— lokhu ngokunye okulakho ukuveza ukuthi ngumbiko olungiswe nge-AI. Kulakho ukuba sengathi yinhlekisa kodwa ukukhangela ama- watermark emifanekisweni lakumavidiyo kulakho ukwahlukanisa imibiko yamanga leyeqiniso. Nxa usecingweni olulevidiyo ucabanga ukuthi yi-deepfake elungiswe ngeAI, tshela umuntu oxoxa laye ukuthi abhodise isihlalo sakhe, nxa engehluleka ukwenza lokhu, kulakho ukuba lucingo lwamanga.”
UCohn uthi i-AI ilakho ukusetshenziswa kuhle kwezombusazwe. Iyenelisa ukuphathisa abakhankasayo ukuthi bakhulume labavoti ngendimi zabo njalo ilakho ukupha abakhankasayo ulwazi ngabavoti ukuze benelise ukubumba imibiko yabo.
Forum