Ama-oda wami we-Windows e-Isitolo se-Windows ubuhlakani bokwenziwa sekuyisihloko sengxoxo ezindaweni eziningi. Kusukela ekukhiqizeni emsebenzini kuya ekwakhiweni kokuqukethwe kwe-multimedia. Kunamasu nezigaba okufanele zicatshangelwe ukuze kukhonjwe amavidiyo noma izithombe ezidalwe kusetshenziswa i-AI. Kulesi sihloko sihlola ezinye ezaziwa kakhulu, nokuthi unganaka kanjani imininingwane nezinkinga ezithile ukuze ugweme ama-deepfakes.
Okokuqala kufanele uqonde ukuthi igama elithi deepfake lisho ukuthini, ukuthi zidalwe kanjani nokuthi zingaba namuphi umthelela. Amavidiyo noma izithombe ezidalwe yi-AI zingabonakala zingokoqobo ngokwedlulele, kodwa ukusetshenziswa kabi kungaholela ngisho nasekungqubuzaneni kwezomthetho.
Amavidiyo noma izithombe ezidalwe yi-AI, yini i-deepfake?
I-Los Ama-Deepfake amavidiyo mbumbulu nezithombe ezidalwe ngama-algorithms ayinkimbinkimbi e-Artificial Intelligence.. Zisetshenziselwa ukwenza kubonakale sengathi umuntu wenza noma usho izinto ezingenzeki ngempela. Ukuthuthuka kwe-Artificial Intelligence kwenze kwaba nzima kakhulu ukuthola ividiyo mbumbulu. Okudaliwe kucishe kungabonakali, kodwa akhona amaqhinga okubona amavidiyo noma izithombe nge-AI angakusiza. Uhlale ucabanga ukuthi akufanele ukholwe yonke into oyibona ezinkundleni zokuxhumana.
Isimo sobuso; izindlebe neminwe
Isici sokuqala okufanele sisinakekele ukuze sizame khomba amavidiyo angamanga ngubuso. Ukuma kobuso futhi ikakhulukazi izindlebe kungakunikeza umkhondo wokuthi ividiyo iyimbumbulu. Lokhu kungenxa ye-asymmetry yobuso, eyinkimbinkimbi kakhulu ukulingisa ngokuphelele ukusebenzisa i-AI. Kuyinkomba enkulu kakhulu yokukhohlisa futhi ingabonwa ikakhulukazi eminwe nasezindlebeni.
UChris Ume, umdali odumile wezithombe esebenzisa i-Artificial Intelligence, utshele isiteshi se-YouTube se-Corridor Crew ukuthi ukwakheka kobuso kuveza imikhondo njalo. Iphuzu elikhulu lokungaboni ngaso linye ezibalweni ze-IA ngokuvamile izindlebe neminwe. Uma ungakwazi ukuqhathanisa izithombe zangempela namavidiyo noma izithombe ezijulile, naka le mininingwane.
ukwehluleka okubukwayo
Ngenkathi i izithombe namavidiyo adalwe nge-AI Angase abukeke ekholisa ngokwedlulele, kukhona amanye ama-engeli lapho kuvela khona amaphutha. Into evame kakhulu ukuthi ziyabonakala lapho siphendulela isithombe ohlangothini olulodwa. Inhlangano ezimele i-Full Fact izinikele ekuqinisekiseni idatha e-United Kingdom futhi incoma “ukwehliswa kwevidiyo ukuze kutholwe amaphutha angaba khona kubuchwepheshe be-AI.” Lokhu kungenxa yokuthi kunamaphuzu athile nezikhathi lapho ubuchwepheshe be-Artificial Intelligence bungahambelani ngokugcwele nengxenye yangempela yevidiyo.
Izinkulumo nokuthinta okudalwe yi-AI kumavidiyo noma ezithombeni
Esinye isihluthulelo soku thola izithombe eziguqulwe i-AI Ziyizinkulumo nokushukuma komzimba. Lolu hlobo lolwazi lwaziwa ngokuthi yi-soft biometrics futhi lungasiza ekutholeni ukungahambisani kwabantu. Kumavidiyo nezithombe ezimayelana nabantu abadumile, kuvame kakhulu ukuthola lokhu kungafani ekusebenzeni kwamathuluzi e-AI.
Ubuchwepheshe be-"Deepfake" busenenkinga yokuthola lezi zimpawu. Yingakho ukungahambisani noma amaphuzu ngaphandle kokuxhumana ngokuvamile kutholakala lapho. Kuyisigaba esithokozisayo sokuhlaziya nokuzama ukuthola ukuthi iyividiyo yangempela noma cha.
ukuqondanisa kwamehlo
Ukunaka indlela amehlo aqondaniswe ngayo nakho kungaba usizo kakhulu ekutholeni imizamo ukukhohlisa nge-Artificial Intelligence. Zama ukusondeza nokumisa ividiyo ngezikhathi ezihlukene. Uma amehlo engagxili ohlangothini olufanayo, kungase kube ushintsho ngenxa ye-Artificial Intelligence. Ukuvumelanisa iso kungenye imininingwane lapho amaphutha abonwa khona. Ubuchwepheshe be-AI okwamanje abukwazi ukugcina unomphela lokhu kuvumelanisa komzimba womuntu. Ngokuka-Ume, lokhu kulula ukukubona kumafreyimu amile.
Ukuvumelanisa izwi nevidiyo
Ekugcineni, isignali eyisihluthulelo khomba ama-deepfakes Ukuvumelanisa okulungile phakathi kwezwi nevidiyo. Lokhu kungenxa yokuma okutholwa ngumlomo lapho siphimisa imisindo ethile. Ubuchwepheshe be-Artificial Intelligence ngokuvamile buba nezinkinga zokuvumelanisa lezi zinhlobo zemisindo. Uma kuziwa ekutholeni ukuthi ividiyo noma isithombe senziwa ngokwedijithali, leli ithuluzi elihle.
Iziphetho
Ngezikhathi ze ngemuva kweqiniso, ukuthola ukuthi ividiyo ingeyangempela yini noma ishintshiwe kubalulekile. Uma ubona isithombe noma ividiyo futhi unokungabaza, khumbula ukusesha eminye imithombo futhi uhlole ukuthi akhona yini amashothi asuka kwamanye ama-engeli. I-Internet iyithuluzi elihle, kodwa kukhona abasakaza imininingwane engamanga bezama ukonakalisa isithunzi sabantu noma bamane benze amahlaya. Kungakho kufanele sicabangele ezinye izinyathelo ezisivumela ukuthi siqinisekise ubuqiniso besithombe noma ividiyo. I-Artificial Intelligence, uma kungenjalo, igcina isiyithuluzi eliyingozi kuye ngokusetshenziswa abasebenzisi abanquma ukukunikeza. Ngalawa masu ungaba ukulungele kancane ukuthola ama-deepfakes.