- Izincomo ezingaphezu kuka-20% kubasebenzisi abasha ku-YouTube zingamavidiyo akhiqizwe yi-AI esezingeni eliphansi
- ISpain ihamba phambili ngenani lababhalisile beziteshi ze-“AI slop”, ezingaphezu kwezigidi ezingu-20.
- Imodeli yesincomo se-YouTube kanye nokwenza imali kukhuthaza ukukhiqizwa okukhulu kokuqukethwe okungenamsoco okuzenzakalelayo
- Lokhu kuguquguquka kwenza kube nzima ukuhlukanisa okuqukethwe kwangempela kokuqukethwe okukhiqizwe yi-AI futhi kwehlisa inani lolwazi lwepulatifomu.
Ezinyangeni zakamuva, abasebenzisi abaningi sebeqalile ukuqaphela ukuthi Okuphakelayo kwe-YouTube, ikakhulukazi kuma-Shorts, sekugcwele amavidiyo angajwayelekile, aphindaphindwayo, futhi acishe abe yize.Akuyona nje imizwa: ukuhlaziywa okuningana okuzimele kubonisa ukuthi ingxenye enkulu yalokho ipulatifomu ekuncomayo kubanikazi bama-akhawunti amasha ayisakhiqizwa ngabantu, kodwa yizinhlelo zobuhlakani bokwenziwa ezikhiqiza amavidiyo ngokulandelana.
Umbiko wakamuva ovela kuwebhusayithi iKapwing ubeka izinombolo kulesi simo futhi usibiza ngokuthi “i-AI slop”: Inani elikhulu lokuqukethwe okukhiqizwe yi-AI, okunenani elincane lokudala futhi okuhloselwe ukujabulisa i-algorithm kuphelaUcwaningo luphetha ngokuthi Amavidiyo angaphezu kuka-20% okuqala angu-500 aphakanyiselwe umsebenzisi omusha asevele ewudoti ozenzakalelayo.Futhi ukuthi amazwe afana neSpain abe ngenye yezindawo eziyinhloko zalolu hlobo lwesiteshi.
Kuyini "i-AI slop" futhi kungani ihlasela i-YouTube?
Igama elithi “slop” selidumile kakhulu kangangokuthi I-Merriam-Webster ikubeka njengegama elifanayo nokuqukethwe kwedijithali okusezingeni eliphansi okwenziwe ngobuningi obukhulu kusetshenziswa i-AINgomongo we-YouTube, lawa amavidiyo akhiqizwe cishe ngaphandle kokuqashwa ngumuntu, ngokusekelwe eziqeshaneni ezikhiqizwe ngu... amamodeli afana neSora noma i-VEO, amazwi okwenziwa kanye nezikripthi ezenziwe ngaphambilini eziphinde zisetshenziswe ngokuphindaphindiwe.
Lawa mavidiyo Azihlose ukunikeza ulwazi, ukuxoxa izindaba zokuqala, noma ukuzama ngobuciko.Inhloso yayo iwukukhohlisa kakhulu: Khulisa ukugcinwa kwezilaleli ngezisusa ezidlulele, imigqi ephindaphindayo, kanye nezindaba ezilula kuze kube sezingeni lokungacabangi, umqondo ofana kakhulu nowokuqukethwe "kokubola kwengqondo" okwasakazeka ezinkundleni zokuxhumana kusukela ngo-2025 kuqhubeke.
Kulolu cwaningo olusekela umbiko, uKapwing wadala amaphrofayili amasha ukuze alingise i-algorithm "ehlanzekile" futhi wahlaziya izincomo zokuqala ezingu-500 ze-YouTube. Cishe ividiyo eyodwa kwezinhlanu ivele yakhiqizwa yi-AIkuyilapho ingxenye yesithathu eyengeziwe iwele ngokuphelele esigabeni sokuqukethwe okungasho lutho nokukhuthazwe ngokweqile okwenzelwe ukugcina umsebenzisi eswayipha ngaphandle kokucabanga kakhulu.
Okubalulekile wukuthi Ukukhiqiza lokhu kushibhile kakhulu futhi cishe akunamkhawulo.Mane nje uhlanganise umbhalo ojwayelekile, i-video generator, ithuluzi lokuhlanganisa inkulumo, kanye namathempulethi ambalwa alungiselelwe i-algorithm. Umphumela uba izinkulungwane zeziqeshana ezibukeka zifana kakhulu ezilayishwa ngokuzenzakalelayo, sinethemba lokuthi uhlelo lokuncoma luzogcina luvuza abambalwa.
ISpain ihola phambili ohlwini lwemfucuza ekhiqizwa yi-AI
Esinye seziphetho ezihlaba umxhwele kakhulu embikweni ukuthi ISpain ihamba phambili emhlabeni wonke ekulinganisweni kwababhalisile beziteshi ezihlukaniswe ngokuthi “i-AI slop” ku-YouTubeUma sihlanganisa izethameli zalolu hlobo lweziteshi, izwe lizodlula ababhalisile abayizigidi ezingu-20, ngaphezu kwe-United States (cishe izigidi ezingu-14) kanye ne-Brazil (cishe izigidi ezingu-13).
Le datha itholakale ngokuhlaziya Iziteshi eziyi-100 ezivele kaningi ezihlokweni ezithrendayo ezweni ngalinyeKwahlaziywa iziteshi ezingaba ngu-15.000. Kulezi, ezingaba ngu-278 zahlukaniswa ngokuthi "i-pure slop," okusho okuqukethwe okukhiqizwe cishe ngokuphelele yi-AI futhi kunephethini ehlelekile yokuphindaphinda, ukweqisa, kanye nenani elincane lolwazi noma lemfundo.
Umbiko ukhomba imakethe yabantu abakhuluma iSpanishi Iziteshi eziningana eziqondiswe kubabukeli baseSpain nabaseLatin America ziphakathi kwezivelele kakhuluAbanye bazethula njengeziteshi zezingane, abanye njengokuzijabulisa okulula noma okuqukethwe okufundiswayo, kodwa bonke babelana ngephethini efanayo: Amavidiyo amafushane, aheha kakhulu akhiqizwe ngesivinini esiphezulu nge-AI.
Isibonelo esimele kakhulu yileso seziteshi zezindaba nezithombe eziphefumulelwe yizindawo ezidumile njengeDragon Ball noma ezinye izinhlaka ezaziwayo. Lezi ziteshi ziqoqa izigidi zabalandeli kanye nezigidigidi zokubukwa zisebenzisa izinhlamvu nezitayela ezivezwa yi-AI ezibonakalayo, kuxutshwe nemidwebo elula kanye nobuhle okuthi, uma uklwebha ubuso, buveze isitembu esizenzakalelayo.
La mavidiyo kadoti enziwa kanjani nge-AI?
Inqubo yokukhiqiza lolu hlobo lokuqukethwe iya ngokuya ijwayelekile. Abadali abaningi - noma kunalokho opharetha beziteshi - baqala ngemibhalo esesidlangalaleninjengezindaba zakudala, izinganekwane, noma izinganekwane. Ukusuka lapho, baphakela ithuluzi lokukhiqiza amavidiyo ukuze bathole izigcawu ezinyakazayo, bengeze ukulandisa okuzenzakalelayo nomculo wangemuva, futhi ngemizuzu embalwa nje bane-clip entsha elungele ukulayishwa.
Ngokusho kwabaphenyi, Amamodeli evidiyo akhiqizayo akuvumela ukuthi udale izigcawu eziphelele ngokuchofoza okumbalwa nje.Ngakho-ke, inkinga ayisekho ekukhiqizweni, kodwa ikhono lokulayisha nokuthuthukisa amavidiyo. Iziteshi eziningi zize zizenzele ngokuzenzakalela le ngxenye, zihlela ukulayisha njalo futhi zisebenzisa izihloko nezithonjana ezenzelwe kuphela ukuheha inani elikhulu lokuchofoza.
Umphumela uyi- Isiphepho esiphelele sokwenza izinto ngokuzenzakalela, ukwenza ngcono i-algorithm, kanye nokuwohloka kwenani lemfundo nolwaziEkuhlolweni ngama-akhawunti amasha, kulinganiselwa ukuthi phakathi kuka-21% no-33% wamavidiyo okuqala angu-500 anconyiwe awela ngaphansi kwelebula elithi "slop", kungaba ngoba akhiqizwa ngqo yi-AI, noma ngoba aphinda umqondo wayo wokugcwala kanye nokukhuthaza okuqhubekayo.
Konke lokhu, ngaphezu kwalokho, kuvame ukwenziwa ngaphandle kwamalebula acacile abonisa ukuthi okuqukethwe kukhiqizwe yi-AIIzixwayiso, uma zikhona, ziyimfihlo noma azikho nhlobo, ngakho abasebenzisi abaningi abakwazi ukuhlukanisa ukuthi kukhona umdali ongumuntu ngemuva kwazo noma uchungechunge lokukhiqiza amavidiyo oluzenzakalelayo.
Imali eningi isengozini futhi inzalo encane ekuyimiseni.
Ngemuva kwalokhu kuchitheka kwemfucumfucu ekhiqizwa yi-AI kunesizathu esisobala: Ibhizinisi liyasebenzaEzinye zeziteshi okukhulunywe ngazo embikweni, ikakhulukazi ezimakethe ezinababukeli abaningi, zingafinyelela ukukhiqiza izigidi eziningana zama-euro ngonyaka ngenzuzo yokukhangisaikakhulukazi uma behlanganisa ukwenza imali ngqo ku-YouTube namadili omkhiqizo kanye nezinye izindlela zebhizinisi.
Kucashunwe amacala afana neziteshi ezithile zezithombe zopopayi zezingane noma imindeni enokubukwa okuqongelelwe okungaphezu kwezigidigidi ezimbili, okukhona Imali engenayo elinganiselwayo idlula kalula izigidi zamaRandi ngonyakaNgisho noma okuqukethwe kukubi futhi kuphindaphinda, uma nje kudala ukubukwa futhi kugcina umsebenzisi epulatifomu, imodeli yebhizinisi iyasebenza.
Lokhu kuchaza ukuthi kungani Kubonakala sengathi akukho sinyathelo esiqondile esivela ku-YouTube sokuvimbela ukusabalala kwalesi simo.Uma nje amavidiyo ehlonipha iziqondiso zomphakathi eziyisisekelo—angavuseleli inzondo, angafaki ubudlova obucacile, njll.—angaqhubeka nokwenza imali futhi, ezimweni eziningi, athuthukiswe uhlelo lokuncoma ngokwalo.
Ngakho-ke ipulatifomu ibhekene nokuphikisana okuyisisekelo: Ukunciphisa "ukwehla kwe-AI" kungasho ukuyeka ingxenye enkulu yemali engenayo kanye nesikhathi sokubuka sonke.Futhi, okwamanje, ibhalansi ibonakala ithambekele kakhulu ekuvumeleni imakethe yemfucuza ezenzakalelayo izilawule yona kunokwethula izihlungi eziqinile ezibeka phambili ikhwalithi kanye nokucaca.
Inkinga akuyona i-AI, yizincomo kanye nemodeli yebhizinisi
Iningi lochwepheshe abaxoxiswene nabo liyavuma ukuthi Ubuhlakani bokwenziwa abuyona isitha.Amathuluzi afanayo asetshenziswa namuhla ukugcwalisa i-YouTube ngamavidiyo angenalutho angasetshenziswa, ezandleni zabadali abanomthwalo wemfanelo, ukuzama amafomethi amasha, ukuhumusha okuqukethwe, noma ukwenza lula ukukhiqizwa okulalelwayo nokubukwayo okuzimele.
Bakhomba ukuthi kufanele kugxilwe kakhulu ku- isincomo kanye nemodeli yokwenza imali yesikhulumi uqoboI-algorithm ibeka phambili inani, imvamisa yokushicilelwa, kanye nekhono lokubamba ukunaka ngaphezu kwezinye izici ezifana nokwangempela, ikhwalithi yolwazi, noma umongo. Kuleyo ndawo, Ukukhiqizwa ngobuningi kwamavidiyo akhiqizwe yi-AI kuhambisana kahle negilavu..
Ezimakethe ezifana neSpain, lapho ividiyo efanayo ngesiSpanishi ingasabalala khona ngesikhathi esisodwa emazweni amaningana, umphumela uphindaphindwaIsiteshi esisodwa "se-AI slop", esilungiswe kahle ngokwe-algorithm, singakha izithameli ezinkulu eSpain naseLatin America ngasikhathi sinye, siqinise kakhulu izisusa zokuqhubeka nokukhiqiza udoti ozenzakalelayo.
Uma i-YouTube ingalethi izinguquko eziyisisekelo endleleni enquma ngayo ukuthi yini ezoyincoma nokuthi isabalalisa kanjani ukwenza imali, Ingozi iwukuthi ipulatifomu izogcina igcwele okuqukethwe okungenalutho.Kuzoba nzima kakhulu ukuthola amavidiyo adalwe abantu abanemibono yabo, izindlela zokuqala, noma ubukhali obunolwazi, ikakhulukazi kwabaqalayo abathembele cishe ngokuphelele eziphakamisweni zokuqala zesistimu.
Ngokombono womsebenzisi, lokhu kuveza isimo lapho Ikhono lokuhlukanisa phakathi kokuqukethwe kwangempela kanye nezingcezu ezikhiqizwe ngobuningi ze-AI liba yinto ebalulekileUkuchichima kwamavidiyo abonakala sengathi angabantu, kodwa angebona, kungaqeda ukwethembana kulokho okubonwayo nokuzwiwayo, futhi kwenze kube nzima nakakhulu uhlelo lolwazi oluseluncane kakade kwi-inthanethi.
Njengoba kunikezwe idatha ekhomba iSpain njengehamba phambili ekubhaliseleni iziteshi ze-“AI slop”, kanye nenani lamavidiyo azenzakalelayo angena ezincomweni zabasebenzisi abasha, konke kukhomba iqiniso lokuthi Udoti odalwe yi-AI ku-YouTube usushintshe ekubeni yinto engavamile waba yingxenye yesakhiwo sendawo yesikhulumi.Uma nje imodeli yebhizinisi iqhubeka nokuvuza inani nokugcinwa kwayo ngaphezu kwekhwalithi kanye nokucaca, futhi kungekho zinyathelo ezicacile zokuhlonza ngokusobala ukuthi yini edalwe nge-AI, inselele kulabo abafuna okuqukethwe okuthembekile - kokubili eSpain nakulo lonke elaseYurophu - kuzoba ukufunda ukuhamba endaweni egcwele amavidiyo acwebezelayo ngaphandle kodwa angenalutho ngaphakathi.
