Kuyiqiniso: izinkampani zase-Brazil zifake i-Artificial Intelligence kumasu azo ebhizinisi—okungenani u-98% wawo, ngokocwaningo olwenziwa ekupheleni kuka-2024. Inkinga, nokho, ukuthi zingama-25% kuphela ezinhlangano ezimemezele ukuthi zikulungele ukusebenzisa i-AI. Abasele bahlushwa ukulinganiselwa kwengqalasizinda, ukuphathwa kwedatha, nokushoda kwamathalente akhethekile. Kodwa lokhu akusho ukuthi ama-75% asele alindele izimo ezikahle zokuqhubekisela phambili amaphrojekthi abo: ngokuphambene nalokho, lezi zinkampani ziyaqhubeka nokusebenzisa ubuchwepheshe.
Inkinga iwukuthi inkampani eyodwa kweziyisihlanu kuphela ekwazi ukuhlanganisa i-AI ebhizinisini layo—ngokombiko osanda kukhululwa womhlaba wonke olungiswe yi-Qlik ngokubambisana ne-ESG. Ngaphezu kwalokho, zingama-47% kuphela izinkampani ezibike ukuthi zisebenzisa izinqubomgomo zokulawula idatha. Lezi zibalo zisemhlabeni wonke—futhi bekungeke kumangaze ukube izibalo zaseBrazil bezingaphezulu kakhulu. Futhi noma i-AI okwamanje isetshenziswa kuma-silo, futhi "indawo yokungena" yobuchwepheshe imvamisa isevisi yamakhasimende, ubungozi bezezimali, bokulawula, kanye nesithunzi sisekhona.
Izinkampani ezikhetha ukusebenzisa i-AI ngaphandle kokulungiselela okufanele zibhekana nezithiyo eziningi. Ucwaningo lwezehlakalo lubonise ukuthi ama-algorithms aphethwe kabi angaqhubeza ukuchema noma onakalise ubumfihlo, okuholela ekulimaleni kwesithunzi nezezimali. Ukubusa kwe-AI akuyona nje inkinga yobuchwepheshe, kodwa futhi eyokusebenza kanye nokukhuthala okufanele: ngaphandle kwesu elichazwe kahle, izingozi zikhula ngokuvumelana namathuba—kusuka ekwephuleni ubumfihlo nokusetshenziswa kabi kwedatha kuya ezinqumweni ezizenzakalelayo ezingacacile noma ezichemile ezidala ukungathembani.
Ingcindezi Elawulayo kanye Nokuthobelana: Izisekelo Zokubusa kwe-AI
Isidingo sokusungula ukubusa kwe-AI asizange nje sivele phambili kwezamabhizinisi: imithethonqubo emisha iyavela, futhi inqubekelaphambili ibishesha, kuhlanganise naseBrazil.
Ngo-December 2024, i-Federal Senate yagunyaza uMthethosivivinywa 2338/2023 , ophakamisa uhlaka lokulawula lwe-AI olunezinkombandlela zokusetshenziswa okufanele. Umthethosivivinywa wamukela indlela esekelwe engcupheni , efana naleyo ye-European Union, ehlukanisa izinhlelo ze-AI ngokwamandla azo okulimaza amalungelo ayisisekelo. Izicelo ezibeka engcupheni eyeqile, njengama-algorithms ezikhali ezizimele noma amathuluzi okuqapha abantu abaningi, zizovinjelwa , amasistimu e-AI akhiqizayo kanye nezinjongo ezijwayelekile kuzodingeka ukuthi ahlole ubungozi ngaphambi kokufika emakethe.
Kukhona nezimfuneko zokubonisa izinto obala, isibonelo, ezidinga onjiniyela ukuthi badalule ukuthi basebenzise okuqukethwe okune-copyright uma beqeqesha amamodeli. Ngesikhathi esifanayo, kunezingxoxo mayelana nokunikeza i-National Data Protection Authority (ANPD) indima eyinhloko ekuxhumaniseni ukuphathwa kwe-AI ezweni, isebenzise uhlaka olukhona lokuvikela idatha. Lezi zinyathelo zomthetho zibonisa ukuthi izinkampani maduze zizoba nezibopho ezicacile mayelana nokuthuthukiswa nokusetshenziswa kwe-AI—kusukela ezinkambisweni zokubika kanye nokunciphisa ubungozi kuya ekubalweni kwemithelela ye-algorithmic.
E-United States naseYurophu, abalawuli bakhulise ukucutshungulwa kwama-algorithms, ikakhulukazi ngemva kokuduma kwamathuluzi e-AI akhiqizayo, okuqubule inkulumompikiswano yomphakathi. I-AI ACT isivele iqale ukusebenza e-EU, futhi ukuqaliswa kwayo kuhlelelwe ukuphela ngo-Agasti 2, 2026, lapho izibopho eziningi zezinga sezisebenza, okuhlanganisa nezidingo zezinhlelo ze-AI eziyingozi kakhulu kanye namamodeli e-AI enhloso evamile.
Ukungafihli, izimiso zokuziphatha kanye nokuziphendulela kwe-algorithmic
Ngale kwesici sezomthetho, ukuphatha kwe-AI kuhlanganisa izimiso zokuziphatha kanye nokuzibophezela okudlula "ukuhambisana nomthetho." Izinkampani ziyaqaphela ukuthi, ukuze zithole ukwethenjwa amakhasimende, abatshalizimali, kanye nomphakathi wonkana, obala mayelana nendlela i-AI esetshenziswa ngayo kubalulekile. Lokhu kuhlanganisa ukwamukela uchungechunge lwezinqubo zangaphakathi, ezinjengokuhlolwa kwangaphambili komthelela we-algorithmic, ukuphathwa okuqinile kwekhwalithi yedatha, nokuhlola okuyimodeli okuzimele.
Kubalulekile futhi ukusebenzisa izinqubomgomo zokulawula idatha ezihlunga ngokucophelela futhi zikhethe idatha yokuqeqeshwa, kugwenywe ukuchema okucwasayo okungase kugxiliswe olwazini oluqoqiwe.
Uma imodeli ye-AI isisebenza, inkampani kufanele ihlole ngezikhathi ezithile, iqinisekise, futhi ihlole ama-algorithms ayo, ibhale izinqumo kanye nenqubo esetshenzisiwe. Leli rekhodi linezinzuzo ezimbili: lisiza ukuchaza ukuthi uhlelo lusebenza kanjani futhi lunike amandla ukuziphendulela uma kwenzeka ukwehluleka noma umphumela ongalungile.
Ukuphatha: ukuqamba okusha okunevelu yokuncintisana
Umbono oyiphutha ojwayelekile ukuthi ukubusa kwe-AI kubeka imingcele emisha. Ngokuphambene nalokho, isu lokuphatha elilungile livumela ukusungulwa okusha okuphephile, ukuvula amandla aphelele e-AI ngokuzibophezela. Izinkampani ezakha izinhlaka zazo zokuphatha kusenesikhathi zinganciphisa izingozi ngaphambi kokuba zibe izinkinga, zigweme ukusetshenzwa kabusha noma amahlazo angabambezela amaphrojekthi.
Ngenxa yalokho, lezi zinhlangano zithola inani elikhulu ngokushesha ezinhlelweni zazo. Ubufakazi bemakethe buqinisa lokhu kuhlobana: inhlolovo yomhlaba wonke yathola ukuthi izinkampani ezinobuholi obusebenzayo bokungamela i-AI zibika imithelela ephakeme yezezimali evela ekusetshenzisweni kwe-AI ethuthukisiwe.
Ngaphezu kwalokho, sisesikhathini lapho abathengi nabatshalizimali beya ngokuya beqaphela ukusetshenziswa kobuchwepheshe ngokuziphatha - futhi ukukhombisa lokhu kuzibophezela ekubuseni kungahlukanisa inkampani emqhudelwaneni.
Ngokwezinto ezibonakalayo, izinhlangano ezinokuphatha okuvuthiwe zibika ukuthuthukiswa hhayi kuphela kwezokuphepha kodwa nasekuthuthukiseni ukusebenza kahle - abaphathi bakhomba ukuncipha kwesikhathi somjikelezo wephrojekthi ye-AI ngenxa yezindinganiso ezicacile kusukela ekuqaleni. Okusho ukuthi, lapho ubumfihlo, ukuchazwa, kanye nezidingo zekhwalithi kucatshangelwa kusenesikhathi esigabeni sokuklama, ukulungisa okumba eqolo kuyagwenywa ngokuhamba kwesikhathi.
Khona-ke, ukubusa kusebenza njengomhlahlandlela wokuqamba okusha okusimeme, okuqondisayo lapho kufanele kutshalwe khona izimali kanye nendlela yokukala izixazululo ngokuzibophezela. Futhi ngokuvumelanisa izinhlelo ze-AI nesu lebhizinisi lenkampani kanye namagugu, ukubusa kuqinisekisa ukuthi ukuqamba okusha njalo kufeza izinjongo zebhizinisi elikhulu nesithunzi, kunokulandela indlela engayodwa noma engase ibe yingozi.
Ukwakhiwa kwesu lokuphatha le-AI, ngaphezu kwakho konke, kuyisu lesu lokubeka endaweni yokuncintisana. Kuhlelo lwanamuhla lwe-ecosystem, lapho amazwe nezinkampani zivaleleke emjahweni wezobuchwepheshe, labo abasungula izinto ezintsha ngokuzethemba nokwethembeka bahola indlela. Izinkampani ezinkulu ezisungula izinhlelo zokuphatha eziphumelelayo ziyakwazi ukulinganisa ukuncishiswa kwengozi nokwandisa izinzuzo ze-AI, kunokuba zidele enye ngenxa yezinye.
Okokugcina, ukubusa kwe-AI akusakhetheki kodwa kuyisu elibalulekile. Ezinkampanini ezinkulu, ukudala isu lokuphatha manje kusho ukuchaza amazinga, izilawuli, kanye namanani azoqondisa ukusetshenziswa kobuhlakani bokwenziwa eminyakeni ezayo. Lokhu kuhlanganisa yonke into kusukela ekuhambisaneni nemithetho ekhulayo kuya ekudaleni izimiso zokuziphatha zangaphakathi kanye nezindlela zokubeka izinto obala, okuhloswe ngazo ukunciphisa ubungozi kanye nokukhulisa inani ngendlela elinganiselayo. Labo abenza izinto ngokushesha bazovuna imivuzo emisha engaguquki kanye nedumela eliqinile, bazibeka phambili emakethe eqhutshwa i-AI eqhubekayo.
