UMthetho we-AI ujolise ekulungiseni eyona miba icinezelayo ye-AI, kubandakanywa nokuthath' icala, ubumfihlo kunye nolwazi olungelulo.
Abaqulunqi bomthetho kwi-EU namhlanje (nge-13 kaMatshi) bamkele uMthetho we-AI obalulekileyo we-bloc ngesininzi esikhulu, ephawula inyathelo elilandelayo lokunyanzelisa eyona mithetho ibanzi yehlabathi kwi-AI.
Umthetho omtsha uya kuchaphazela indlela iteknoloji ekhula ngokukhawuleza elawulwa ngayo kwihlabathi laseNtshona, ngenxa yokunqongophala kwayo nayiphi na i-guardrails esemthethweni eyenziwa yi-US.
UMthetho we-AI ujolise ekulungiseni eyona miba icinezelayo ye-AI: kubandakanywa ukuthath' icala, ubumfihlo kunye nolwazi olungelulo.
Umthetho uya kumisela amanqanaba omngcipheko ahlukeneyo kwizicelo kunye neenkqubo ze-AI zoluntu. Abo banamanqanaba aphezulu omngcipheko baya kunyanzeliswa ukuba bahlangabezane neemfuno ezingqongqo zokusebenza.
Esinye isikhokelo esisinyanzelo kwiinkqubo ze-AI siya kuba luvavanyo lwamalungelo oluntu olunokulinganisa ukuba inkqubo inocalucalulo okanye inocalucalulo.
Izithintelo ziya kunyanzeliswa ngamaxesha athile ukuvumela ababoneleli be-AI ixesha lokulungelelanisa. Ubuncinci i-15% yazo zonke iinkqubo ze-AI ziya kuwela phantsi kwecandelo lomngcipheko ophezulu.
Iseti yokuqala yezithintelo iya kumiselwa kwiinkqubo zeGenAI ezifana ne-OpenAI's ChatGPT kunye neGemini kaGoogle. Ezinye izithintelo azinakunyanzeliswa kude kube ngu-2026.
"IYurophu ngoku i-standard standard-setter kwi-AI ethembekileyo," kusho uThierry Breton, uMkomishinala weMarike yaNgaphakathi we-EU, kwingxelo.
Ngexesha lentlanganiso yangoLwesibini, abenzi bomthetho be-EU baqinisekisile ukuba mninzi umsebenzi osele wenziwe ngaphambi kokuba lo Mthetho usebenze.
UDragos Tudorache, ilungu lePalamente yaseYurophu, uthe: "Imithetho esiyipasile kulo myalelo wokulawula i-digital domain - kungekhona nje uMthetho we-AI - iyimbali ngokwenene, ubuvulindlela.
Kodwa ukuzenza zonke zisebenze ngokuhambelana nesiphumo esifunekayo kunye nokuguqula iYurophu ibe yindawo yamandla edijithali yekamva iya kuba luvavanyo lobomi bethu.
Uza kuwachaphazela njani amashishini uMthetho we-EU AI?
ULaura Petrone, umhlalutyi oyintloko kwinkampani yophando kunye nohlalutyo lwe-GlobalData, uxelele I-Verdict ukuba uMthetho we-AI uya kuba yingxaki kumashishini athile e-EU.
UPetrone uthe amashishini amaninzi angawufumana lo mthetho “unzima kakhulu” kwaye uthi “ubeka emngciphekweni wokuthintela izinto ezintsha.”
Ukubandakanywa kwemithetho ethintelayo kwiimodeli ezisisiseko, njenge-ChatGPT, kuthetha ukuba ababoneleli baya kunyanzeliswa ukuba balandele iimfuno ezicacileyo ezicacileyo.
"Abaphuhlisi beemodeli ezijoliswe ngokubanzi ezibhalwe njengomngcipheko wenkqubo kuya kufuneka babeke izicwangciso zokunciphisa kwaye badlulisele iinkcukacha zaso nasiphi na isiganeko kwi-AI ye-AI ye-AI ye-European Office, ejongene nokunyanzelisa imithetho," wongezelela uPetrone.
UPetrone uthe iya kuba "yingxaki kodwa ingumceli mngeni" kwi-EU ukwahlula phakathi kweentlobo zeemodeli ezisisiseko kunye nezo zinokufuna uqwalaselo olongezelelweyo lolawulo.
"Kusekho ukungaqiniseki okuninzi malunga nendlela oza kuphunyezwa ngayo umthetho kunye nokuba i-Ofisi ye-AI inazo izixhobo eziyimfuneko ukuze ifezekise imisebenzi yayo," utshilo uPetrone.
UCurtis Wilson, isazinzulu sedatha yabasebenzi kwi-Synopsys Software Integrity Group, uxelele isigwebo ukuba izikhokelo zokulawula ezifana ne-EU AI Act ziya kunceda ukwakha ukuthembela kwi-AI, into evakalelwa kukuba yeyona ngxaki inkulu ejongene nabaphuhlisi be-AI.
"Imithetho engqongqo kunye nezohlwayo ziya kuthintela abaphuhlisi abangakhathaliyo, kwaye bancede abathengi baqiniseke ngakumbi ekuthembeni nasekusebenziseni iinkqubo ze-AI," watsho.
Umthombo ovela isigwebo
Ukuziphendulela: Ulwazi oluchazwe ngasentla lunikezelwa yi-verdict.co.uk ngaphandle kwe-Cooig.com. I-Cooig.com ayenzi lumelo kunye neziqinisekiso malunga nomgangatho kunye nokuthembeka komthengisi kunye neemveliso.
