Umthetho we-AI uhlose ukubhekana nezinkinga ze-AI ezicindezela kakhulu, ezihlanganisa ukuchema, ubumfihlo kanye nolwazi olunganembile.
Abenzi bomthetho e-EU namuhla (13 Mashi) bagunyaze umthetho oyingqophamlando we-AI Act ngeningi elikhulu, okumaka isinyathelo esilandelayo sokuphoqelela isethi yemithetho ebanzi kakhulu emhlabeni ku-AI.
Lo mthetho omusha uzothinta ukuthi ubuchwepheshe obukhula ngokushesha bulawulwa kanjani emazweni aseNtshonalanga, uma kubhekwa ukushoda kwanoma yiziphi izindlela zokuqapha ezisemthethweni ezisetshenziswa yi-US.
Umthetho we-AI uhlose ukubhekana nezinkinga ze-AI ezicindezela kakhulu: okuhlanganisa ukuchema, ubumfihlo kanye nolwazi olunganembile.
Umthetho uzonquma amazinga ahlukene engcuphe ezinhlelo zokusebenza zomphakathi ze-AI namasistimu. Labo abanamazinga aphezulu obungozi bazophoqeleka ukuthi bahlangabezane nezidingo eziqinile ukuze basebenze.
Umhlahlandlela owodwa oyisibopho wezinhlelo ze-AI kuzoba ukuhlolwa kwamalungelo abantu okungakala ukuthi uhlelo luchemile noma luyabandlulula.
Imikhawulo izosetshenziswa ngezikhathi ezithile ukuze kuvunyelwe abahlinzeki be-AI isikhathi sokulungisa. Okungenani u-15% wawo wonke amasistimu e-AI azowela ngaphansi kwesigaba esinobungozi obukhulu.
Isethi yokuqala yemikhawulo izofakwa ezinhlelweni ze-GenAI ezifana ne-OpenAI's ChatGPT kanye ne-Gemini yakwaGoogle. Eminye imikhawulo akumele isetshenziswe kuze kube ngu-2026.
"I-Europe manje isiyi-standard-setter ye-AI ethembekile," kusho u-Thierry Breton, uKhomishani Wemakethe Yangaphakathi we-EU, esitatimendeni.
Ngesikhathi somhlangano wangoLwesibili, abenzi bezomthetho be-EU baqinisekise ukuthi usemningi umsebenzi osamele wenziwe ngaphambi kokuthi uMthetho uqale futhi usebenze.
UDragos Tudorache, oyilungu lePhalamende laseYurophu, uthe: “Imithetho esiyiphasile kuleli gunya lokuphatha isizinda sedijithali - hhayi nje uMthetho we-AI - ingumlando ngempela, iphayona.
"Kepha ukubenza bonke basebenze ngokuvumelana nomphumela ofisayo futhi ukuguqula iYurophu ibe yindawo enamandla yedijithali yesikhathi esizayo kuzoba uvivinyo lwempilo yethu yonke."
Ngabe umthetho we-EU AI uzowathinta kanjani amabhizinisi?
ULaura Petrone, umhlaziyi oyinhloko enkampanini yocwaningo nokuhlaziya i-GlobalData, utshele i-Verdict ukuthi uMthetho we-AI uzoba yinkinga kwamanye amabhizinisi e-EU.
UPetrone uthe amabhizinisi amaningi azowuthola lo mthetho “unzima kakhulu” futhi athi “ubeka engcupheni yokukhinyabeza izinto ezintsha.”
Ukufakwa kwemithetho evimbelayo kumamodeli ayisisekelo, njenge-ChatGPT, kusho ukuthi abahlinzeki bazophoqeleka ukuthi balandele izimfuneko zokubonisa ngale.
“Abathuthukisi bamamodeli enhloso ejwayelekile abhalwe njengengcuphe yesistimu kuzodingeka futhi babeke amasu okunciphisa endaweni futhi badlulisele imininingwane yanoma yisiphi isigameko eHhovisi le-AI elisha leKhomishini YaseYurophu, elibhekele ukuphoqelela imithetho,” kwengeza uPetrone.
UPetrone uthe kuzoba "okubucayi kodwa okuyinselele" ukuthi i-EU ihlukanise phakathi kwezinhlobo zamamodeli ayisisekelo nalawo angadinga ukucutshungulwa okwengeziwe kokulawula.
“Kusekuningi ukungaqiniseki mayelana nokuthi lo mthetho uzosetshenziswa kanjani nokuthi iHhovisi le-AI linazo yini izinsiza ezidingekayo ukuze lifeze imisebenzi yalo,” kusho uPetrone.
UCurtis Wilson, usosayensi wedatha wabasebenzi eSynopsys Software Integrity Group, utshele isinqumo ukuthi izinhlaka zokulawula ezifana ne-EU AI Act zizosiza ukwakha ukwethenjwa ku-AI, into anomuzwa wokuthi inkinga enkulu kakhulu ebhekene nabathuthukisi be-AI.
"Imithetho eqinile kanye nenhlawulo yokujezisa izovimbela abathuthukisi abanganaki, futhi isize abathengi ukuthi baqiniseke ngokwengeziwe ekuthembeni nasekusebenziseni izinhlelo ze-AI," esho.
Umthombo ovela isinqumo
Umshwana wokuzihlangula: Ulwazi olubekwe ngenhla luhlinzekwa yi-verdict.co.uk ngaphandle kwe-Chovm.com. I-Chovm.com ayenzi izethulo namawaranti mayelana nekhwalithi nokuthembeka komdayisi nemikhiqizo.
