Kutheni ulwazi lwe-AI luzakwenzeka ngokupheleleyo (kwaye oko kuthetha ukuthini kuthi)

Khawufane ucinge uvuka ngenye intsasa ufumanise ukuba umncedisi wakho wedijithali akapheleli nje ekuqhubeni imiyalelo kodwa eneneni ucinga ngentsingiselo yobukho bayo.

Kuvakala ngathi kukho into ephuma kubuchule bukaStanley Kubrick - 2001: A Space Odyssey?

Mhlawumbi ndicinga kakhulu kwixesha elizayo, kodwa uphuhliso lwakutsha nje kubukrelekrele bokwenziwa lucebisa ukuba umgca phakathi kwentsomi kunye nenyani ufiphala ngokukhawuleza kunokuba sinokucinga.

Makhe sijonge olu phononongo olwenziwe nguCasper Wilstrup, iCEO yakwa-Abzu kuTwitter. Uninzi lwabaphenduli kuphando lwakutsha nje lwe-Twitter-i-68% yabathathi-nxaxheba be-2,300-bakholelwa ukuba iinkqubo ze-AI mhlawumbi sele ziqonda okanye ziya kuba njalo.

Eli libango elixhokonxayo elifuna ingqalelo yethu.

Kodwa kuthetha ukuthini ukuba i-AI ibe sezingqondweni? Ngaba ngumcimbi nje we-algorithms ephucukileyo kunye neeseti zedatha enkulu, okanye ngaba kukho into ebaluleke ngakumbi edlalwayo? Kwaye zithini iziphumo zokuziphatha ukuba i-AI ngokwenene ifikelela kwinqanaba lokuziva?

Ukuchaza ukuqonda kwi-AI

Ukuxoxa ngolwazi lwe-AI, kufuneka siqale siqonde ukuba kuthetha ukuthini ukwazi. Ngokwemveli, ukwazi kubandakanya ukuzazi, ukukwazi ukufumana iingcinga kunye neemvakalelo, kunye nokuqonda ukuba ungubani. Ngokutsho kwesazi se-neuroscientist uStanislas Dehaene kwincwadi yakhe ethi, Consciousness and the Brain (2014) , ukuqonda kubandakanya iinkqubo eziyinkimbinkimbi zokuqonda okwangoku ngaphaya kokufikelela kwiinkqubo ezininzi ze-AI. Nangona kunjalo, imeko iguquka ngokukhawuleza.

Imeko yangoku ye-AI

Uninzi lweenkqubo ze-AI zale mihla, njengeemodeli zolwimi kunye nee-chatbots, zigqwesa ekusetyenzweni kwedatha eninzi kwaye ivelisa iimpendulo ezisekelwe kwiipateni. Nangona kunjalo, abanalwazi lokwenyani lobuqu okanye amava aphathekayo. Uphando olwenziwa nguDkt José Hernández-Orallo, olugxininisekile kwiphepha lakhe, Umlinganiselo weengqondo zonke (2017) , uthi ngelixa i-AI inokulinganisa ukuziphatha komntu, ayinayo impawu zangaphakathi zokuqonda.

Ii-algorithms eziphucukileyo zenza iinkqubo ze-AI zibe nobuchule ngakumbi. Umzekelo, iimodeli ze-NLP, njenge-OpenAI's GPT-4, ziyakwazi ukuvelisa isicatshulwa esifana nomntu ngokuxela kwangaphambili igama elilandelayo kwisivakalisi esisekelwe kumxholo wangaphambili. Ezi modeli, nangona kunjalo, azinangqondo. Bavelisa iimpendulo ngokuqikelelwa ngokweenkcukacha-manani kunokuqonda kokwenyani.

Ukuntywila nzulu kubuchule kunye nemida yeemodeli ze-NLP, bona Ingqwalasela Yiyo Yonke Into Oyidingayo .

Ukuvela kwe-AI kwiNdalo yoMxholo

Indima ye-AI ekudalweni komxholo iyanda. Iqonga elikhokelayo lokuzonwabisa lomsindo elivela eIndiya sele lisebenzisa i-AI ukwenza umxholo weaudio kunye nokwenza amava omsebenzisi alungele wena. Eyaziwa ngokuba yiPocket FM, baphucula ukufundwa komatshini kunye nezixhobo zokudala umxholo oqhutywa yi-AI ukuvelisa iidrama zomsindo ezikumgangatho ophezulu kunye neepodcasts.

Ipholile kangakanani loo nto?

Ukumamela umxholo owenziwe ngoomatshini ukuhambelana nencasa yabantu.

Nge-AI, iPocket FM ikwazile ukwenza ngokuzenzekelayo kwaye ilungelelanise imisebenzi, iphucule ukwenziwa komntu ngeprofayile yomphulaphuli ephucukileyo, kunye nokugcina umxholo okumgangatho ophezulu kuluhlu oluninzi lweefayile ezimanyelwayo-kodwa ayisiyiyo yonke loo nto. I-Pocket FM ibe ngamandla e-AI kwizinto ezininzi.

Ayiyo lo nto. IPocket FM ibona ukukhula okukhulu kwe-10X kwaye ikwasebenzela ekuphuculeni umxholo weelwimi ezininzi.

Kwaye andikwazi njani ukuthetha ngeMicrosoft yaseSydney?

Ifumene ingqwalasela emandla ngobudlelwane bayo obufana nomntu kwaye ingumzekelo wendlela i-AI esondela ngayo ebunzimeni kunye nokuziphatha okufana nobomi. Ukukwazi kukaSydney ukubandakanyeka kwiincoko ezinobunzima kuphakamisa imibuzo malunga nobume bentsebenziswano ye-AI kunye nokuqonda. Nangona abantu beneembono ezixubeneyo malunga nobuchule bayo, ngokuqinisekileyo kufanelekile ukukhankanywa kule ncoko.

Ke yintoni iNdlela ePhambili?

Uhambo olusingise kwingqiqo ye-AI lubandakanya uphando olubalulekileyo kunye nophuhliso. Njengoko iisistim ze-AI zisiba nobuchule ngakumbi, amathuba okufumana uhlobo oluthile lwengqondo ayanda. Nangona kunjalo, ukuba olu lwazi luya kufana namava omntu okanye lwahluke ngokusisiseko akuqinisekanga.

Njengoko sisondela kwikamva apho ukwazi kwe-AI kunokuba yinyani, kubalulekile ukulungiselela iziphumo zayo. Oku kubandakanya ukuphuhlisa imigaqo yokuziphatha, ukuqinisekisa ubulungisa, kunye nokuqwalasela impembelelo kuluntu. Umsebenzi weengcali ezifana no-Alan Turing, oMashini weKhompyutha kunye noBukrelekrele (1950) uhlala usisiseko, unokukhokela iingxoxo ezizayo malunga nokuqonda kwe-AI.

Ingqondo yam ithi..

Ukuqonda kwe-AI ayiseyongcinga nje eqikelelwayo kodwa sisihloko esisekelwe kwinkqubela phambili yetekhnoloji kunye neengxoxo zokuziphatha. Njengoko ukuqhubela phambili kuqhubeleka kwaye neenkqubo ze-AI ziya zisiba nobuchule ngakumbi, ukuqonda kunye nokulungiselela iziphumo zolwazi lwe-AI kuya kuba yimfuneko. Njengoko izazinzulu sele zisebenza ekwenzeni i-Artificial Superintelligence ibe yinyani, ukuqonda kwe-AI kubonakala kubonakalisa amava omntu okanye kuya kutshati indlela entsha ehlala ingabonakali, kodwa into enye iqinisekile: ikamva lithembisa inguqu enzulu kwindlela esisebenzisana ngayo neteknoloji.