Njengoba kuvela, yebo, kodwa kancane

I-AI itholakala yonke indawo. Enye yezindawo ezivame kakhulu i-AI ekhaya kulezi zinsuku i-smartphone, amakhaya ahlakaniphile nezinto zikagesi. Muva nje, njengoba sihlezi sidla isidlo sakusihlwa, sibe nengxoxo no-Alexa ehambe kanje:

Me: Alexa, dlala amaphuzu avelele Amantshontsho. [Lesi isici esikhonjiswa esikrinini sasekhaya se-Alexa. Cela i-Alexa ukuthi idlalele iqembu lakho olithandayo amaphuzu avelele.]

Alexa: Ngithole okuthile kuwebhu. [Ngiyazi uma i-Alexa iqala kanje, kunenkinga. Ngeke kuhambe kahle. I-Alexa ibonisa uhlu lwamavidiyo ambalwa. Kuyavunywa, iningi lawo amavidiyo e-baseball anabadlali abenza imidlalo ehlukile kule minyaka engu-5 edlule. Iphutha lami. Zama futhi.]

Me: Alexa, ngibonise amaphuzu avelele omdlalo we-baseball wakamuva we-Chicago Cubs. [Ngethemba ukuthi ngeke ilibone ithoni yokweyisa njengoba ngamukela ukusolwa ngokungakwazi ukuqonda isicelo sami.]

Alexa: Izithombe ezinhle kakhulu zomdlalo we-Chicago Cubs zizotholakala emahoreni amabili ngemuva kokuphela komdlalo. [Inqubekelaphambili. Bengingazi ukuthi kungenzeka ukuthi bayadlala ngalesi sikhathi. Inhlanhla ingakimi. Ngivele ngibe nethemba.]

Me: Alexa, kuhle ngibonise amaphuzu avelele izolo. [Yebo, ukukhungatheka kwami ​​sekuqala ukubonakala. Sengisondele kakhulu ekwephuleni ikhodi. Ngicishe ngiyinambithe.]

Alexa: Ngiyaxolisa, angimazi lowo. [Isho lokhu kaningi kakhulu. Mhlawumbe bengingacacile.]

Mina: Uyadlala? Dlala, izithombe ezinhle kakhulu zevidiyo zomdlalo we-Major League baseball phakathi kwe-Chicago Cubs ne-Pittsburgh Pirates wangoMsombuluko Julayi 25, 2022 e-Wrigley Field. [Kulokhu ngiyazethemba ukuthi sengiyihlabe esikhonkosini. Ngikhiphe isicelo esiqondile, esicacile okuyikhono engaziyo ukuthi i-Alexa inalo. Ike yakwenza lokhu ngaphambili. ]

amawebhusayithi: [Athule. Lutho. Ayikho impendulo. Ngikhohliwe ukusho izwi lokuvusa umlingo, Alexa.]

The Isilinganiso se-IQ oneminyaka engu-18 ubudala cishe uneminyaka eyi-100. I-IQ emaphakathi yomuntu oneminyaka engu-6 ubudala ingama-55. I-Google AI IQ yahlolwa yaba ngu-47. I-IQ ka-Siri ilinganiselwa ku-24. UBing no-Baidu baseminyakeni yama-30. Angizange ngithole ukuhlolwa kwe-Alexa's IQ, kodwa ulwazi lwami lwalufana nokukhuluma nengane yasenkulisa.

Abanye bangase bathi, akulungile ukunikeza ikhompuyutha ukuhlolwa kwe-IQ. Kodwa, lelo yiphuzu ngokuphelele. Isithembiso se-AI siwukwenza lokho abantu abakwenzayo, okungcono kuphela. Kuze kube manje, i-head-to-head ngayinye - noma, singasho, inethiwekhi ye-neural kunethiwekhi ye-neural -challenge igxile kakhulu. Ukudlala i-chess. Ukuhlola izifo. Ukusenga izinkomo. Izimoto zokushayela. Irobhothi ngokuvamile liyawina. Engifuna ukukubona uWatson esenga inkomo eshayela imoto futhi edlala iJeopardy. Manje, ukuthi kungaba trifecta. Abantu abakwazi nokubheka ugwayi wabo ngesikhathi beshayela ngaphandle kokuthola ingozi.

Inani eliphakeme kakhulu lama-AI

Ikhishwe ngomshini. Ngisola ukuthi angingedwa. Ngaqala ukucabanga, uma lesi kuyisimo sobuciko, zihlakaniphe kangakanani lezi zinto? Singaqhathanisa ubuhlakani bomuntu nomshini?

Ososayensi bayahlola amakhono amasistimu okufunda nokucabanga. Kuze kube manje, abantu bokwenziwa abakenzi kahle njengangempela. Abacwaningi basebenzisa ukushiyeka ukukhomba izikhala ukuze siqonde kangcono lapho intuthuko eyengeziwe nenqubekelaphambili idinga ukwenziwa.

Ukuze nje ungaphuthelwa iphuzu futhi ukhohlwe ukuthi u-“I” ku-AI umeleni, abakhangisi manje sebeqambe igama elithi Smart AI.

Ingabe i-AI Inengqondo?

Ingabe amarobhothi anemizwa? Amakhompuyutha angaba nakho emotions? Cha asiqhubeke. Uma ufuna funda ngakho, injini ye-Google eyodwa (yangaphambili) ithi imodeli ye-AI i-Google esebenza ngayo inengqondo. Waba nengxoxo esabekayo ne-bot eyamenza waqiniseka ukuthi ikhompuyutha inemizwa. Ikhompyutha yesabela impilo yayo. Angikholwa nokuthi ngiwubhalile lowo musho. Amakhompiyutha awanakukwesaba ukuphila. Amakhompyutha awakwazi ukucabanga. Ama-algorithms awacatshangwa.

Nokho, ngeke ngimangale uma ikhompuyutha iphendula umyalo maduze nje ngokuthi: “Ngiyaxolisa, Dave, angikwazi ukwenza lokho.”

Ihluleka kuphi i-AI?

Noma, ngokunembile, kungani amaphrojekthi we-AI ehluleka? Bahluleka ngenxa yezizathu ezifanayo amaphrojekthi we-IT ahlala ehluleka. Amaphrojekthi ayahluleka ngenxa yokungaphathwa kahle, noma ukwehluleka ukuphatha isikhathi, ububanzi noma isabelomali..:

• Umbono ongacacile noma ongachazwanga. Isu elibi. Kungenzeka ukuthi uzwile abaphathi bethi, "Sidinga nje ukuthikha ibhokisi." Uma isiphakamiso senani singachazwa, inhloso ayicacile.
• Ukulindela okungenangqondo. Lokhu kungase kube ngenxa yokungaqondi kahle, ukukhulumisana okungekuhle, noma ukuhlela okungenangqondo. Okulindelwe okungenangqondo nakho kungabangelwa ukuntula ukuqonda kwamakhono kanye nendlela yokusebenza yamathuluzi e-AI.
• Izidingo ezingamukelekile. Izidingo zebhizinisi azichazwanga kahle. Amamethrikhi empumelelo awacacile. Futhi kulesi sigaba ukubukelwa phansi kwabasebenzi abaqonda idatha.
• Amaphrojekthi angabelwe futhi abukelwa phansi. Izindleko azikalinganiselwa ngokugcwele nangokuqondile. Izimo eziphuthumayo bezingakahlelelwa futhi azilindelwe. Igalelo lesikhathi labasebenzi asebematasa kakhulu libukelwa phansi.
• Izimo ezingalindelekile. Yebo, ithuba liyenzeka, kodwa ngicabanga ukuthi lokhu kuwela ngaphansi kokuhleleka okungalungile.

Bheka futhi, okuthunyelwe kwethu kwangaphambilini Izizathu eziyi-12 Zokuhluleka Kwezibalo Nobuhlakani Bebhizinisi.

I-AI, namuhla, inamandla kakhulu futhi ingasiza izinkampani ukuba zithole impumelelo enkulu. Uma izinhlelo ze-AI zehluleka, ukwehluleka cishe kungalandelelwa kokukodwa kwalokhu okungenhla.

Ikuphi i-AI Excel?

I-AI inhle emisebenzini ephindaphindayo, eyinkimbinkimbi. (Uma sikhuluma iqiniso, ingenza nemisebenzi elula, engaphindaphindi, futhi. Kodwa, kungaba kushibhe ukwenza ingane yakho yasenkulisa ikwenze.) Kuhle ukuthola amaphethini nobudlelwano, uma bukhona, ngamanani amaningi edatha.

• I-AI yenza kahle uma ibheka imicimbi engafani namaphethini athile.
  • Ukuthola ukukhwabanisa kwekhadi lesikweletu imayelana nokuthola okwenziwayo okungalandeli amaphethini okusetshenziswa. Ivame ukwenza iphutha ohlangothini lokuqapha. Ngithole izingcingo ezivela ekhadini lami lesikweletu nge-algorithm yokushisekela ngokweqile lapho ngigcwalisa imoto yami eqashiwe ngophethiloli e-Dallas ngase ngigcwalisa imoto yami yomuntu siqu e-Chicago. Bekusemthethweni, kodwa kungajwayelekile ukuthi umakwe.

"American Express icubungula u-$1 trillion ekuthengiseni futhi inamakhadi e-AmEx ayizigidi ezingu-110 asebenzayo. Bathembela kakhulu ekuhlaziyweni kwedatha naku-algorithms yokufunda komshini ukusiza ukuthola ukukhwabanisa maduze nje, ngakho-ke konga izigidi ekulahlekeni”.

• Ukukhwabanisa nokuhlukunyezwa kwemithi. Amasistimu angathola amaphethini okuziphatha angajwayelekile ngokusekelwe emithethweni eminingi ehleliwe. Isibonelo, uma isiguli sibona odokotela abathathu abahlukene edolobheni ngosuku olulodwa benezikhalazo ezifanayo zobuhlungu, uphenyo olwengeziwe lungase lugunyazwe ukuze kukhishwe ukuhlukumeza.
• I-AI phakathi ukunakekela impilo ube nempumelelo enhle kakhulu.
  • I-AI nokufunda okujulile kwafundiswa ukuqhathanisa ama-X-ray nokutholakele okuvamile. Ikwazile ukukhuphula umsebenzi we-radiologists ngokuhlaba umkhosi okungavamile ukuze isazi se-radiologist sihlole.
• I-AI isebenza kahle nayo social kanye yezitolo. Esinye sezizathu ezenza sibone lokhu kakhulu ukuthi kunengozi encane. Ingozi yokuthi i-AI ibe nephutha futhi ibe nemiphumela emibi iphansi.
• • Uma uthanda/uthengile lokhu, sicabanga ukuthi uzothanda lokhu. Ukusuka e-Amazon kuye ku-Netflix naku-YouTube, bonke basebenzisa uhlobo oluthile lokuqashelwa kwephethini. I-Instagram AI ibheka ukusebenzisana kwakho ukuze kugxilwe kokuphakelayo kwakho. Lokhu kuvame ukusebenza kangcono uma i-algorithm ingafaka okuncamelayo ebhakedeni noma eqenjini labanye abasebenzisi abenze ukukhetha okufanayo, noma uma izintshisekelo zakho zincane.
  • I-AI ijabulele impumelelo ethile nayo ukuqashelwa kobuso. I-Facebook iyakwazi ukukhomba umuntu omakwe ngaphambilini esithombeni esisha. Amanye amasistimu okubona ubuso ahlobene nokuvikeleka kusenesikhathi akhohliswa imaski.
• I-AI ijabulele impumelelo ku ukulima kusetshenziswa ukufunda ngomshini, izinzwa ze-IoT namasistimu axhunyiwe.
  • I-AI isizile ogandaganda abahlakaniphile ukutshala nokuvuna amasimu ukuze kwandiswe isivuno, kuncishiswe umanyolo futhi kuthuthukiswe izindleko zokukhiqiza ukudla.
  • Ngamaphuzu edatha avela kumamephu e-3-D, izinzwa zomhlabathi, ama-drones, amaphethini wesimo sezulu, agadiwe ukufunda imishini ithola amaphethini kumasethi edatha amakhulu ukuze ibikezele isikhathi esingcono kakhulu sokutshala izitshalo futhi ibikezele izivuno ngaphambi kokuthi zitshalwe.
  • Amapulazi obisi sebenzisa amarobhothi e-AI ukuze abe nobisi lwezinkomo ngokwawo, i-AI nokufunda ngomshini kuphinde kuqaphe izimpawu ezibalulekile zenkomo, umsebenzi, ukudla namanzi athathayo ukuzigcina ziphilile futhi zanelisekile.
  • Ngosizo lwe-AI, abalimi abangaphansi kuka-2% wabantu bondla izigidi ezingama-300 kulo lonke elase-USA.
  • I-Artificial Intelligence Kwezolimo

Kukhona nezindaba ezinhle ze-AI impumelelo ezimbonini zezinsiza, ezidayisayo, abezindaba nezokukhiqiza. I-AI ikhona yonke indawo ngempela.

Amandla e-AI Nobuthakathaka Kuqhathaniswa

Ukuqonda okuqinile kwamandla nobuthakathaka be-AI kungase kube nomthelela empumelelweni yezinhlelo zakho ze-AI. Khumbula futhi ukuthi amakhono okwamanje akukholomu engakwesokudla amathuba. Lezi yizindawo lapho abathengisi kanye nabamukeli be-blood edge benza inqubekela phambili njengamanje. Sizobheka amakhono njengamanje aphonsa inselelo ku-AI futhi ngonyaka futhi sibhale i-shift kwesokunxele. Uma ufunda ishadi elilandelayo ngokucophelela, ngeke ngimangale uma kukhona ukunyakaza phakathi kwesikhathi engibhala ngaso lokhu nesikhathi okushicilelwe ngaso.

Liyini Ikusasa Le-AI?

Ukube i-AI ibihlakaniphile, ibingase ibikezele ukuthi ikusasa liphetheni. Kuyacaca ukuthi baningi imibono eyiphutha mayelana nalokho i-AI ekwazi ukuyenza nengakwazi ukuyenza. Abaningi imibono eyiphutha kanye nokungafundi kwe-AI ziwumphumela wokumaketha kwe-tech over-hyping amakhono akhona. I-AI iyamangalisa engakwenza namuhla. Ngibikezela ukuthi ubuthakathaka obuningi kukholamu engakwesokudla buzoshintshela kwesokunxele bese buba amandla eminyakeni emi-2 noma emi-3 elandelayo.

[Ngemva kokuqeda lesi sihloko, ngethule isigaba esandulele ku I-OpenAI, i-generator yolimi yesikhulumi se-AI evulekile. Kungenzeka ukuthi ububonile obunye ubuciko obukhiqizwe yi-DALL-E yayo. Bengifuna ukwazi ukuthi icabangani ngekusasa le-AI. Nakhu okwakushiwo. ]

Ikusasa le-AI alikona ukuthenga amaseva ambalwa nokufaka iphakheji yesofthiwe engaphandle kweshalofu. Imayelana nokuthola nokuqasha abantu abalungile, ukwakha ithimba elifanele, kanye nokwenza izimali ezifanele kukho kokubili ihadiwe nesofthiwe.

Eminye impumelelo engaba khona ye-AI eminyakeni embalwa ezayo ihlanganisa:

• Ukwandisa ukunemba kwezibikezelo nezincomo
• Ukwenza ngcono izinqubo zokuthatha izinqumo
• Ukusheshisa ucwaningo nentuthuko
• Ukusiza ukwenza ngokuzenzakalelayo kanye nokuthuthukisa izinqubo zebhizinisi

Kodwa-ke, kukhona futhi okunye ukwehluleka kwe-AI okungenzeka ukuthi amabhizinisi kufanele akuqaphele, okufana:

• Ukuthembela ngokweqile ku-AI okuholela ezinqumweni eziphansi
• Ukungaqondi ukuthi i-AI isebenza kanjani okuholela ekusetshenzisweni kabi
• Ukuchema kwedatha esetshenziselwa ukuqeqesha amamodeli e-AI okuholela emiphumeleni engalungile
• Ukuphepha nokukhathazeka okuyimfihlo kudatha esetshenziswa ukuqeqesha amamodeli e-AI

Ngakho-ke, kusho ukuthini lokhu kumabhizinisi atshala imali ku-AI ukuze agcwalise izibalo zawo zendabuko? Impendulo emfushane ithi, azikho izindlela ezinqamulelayo. U-85% wezinhlelo ze-AI ziyehluleka. Kuyathakazelisa ukuthi lokhu kufana nezibalo ezicashunwa kaningi ezihlobene namaphrojekthi endabuko e-IT nawe-BI. Umsebenzi onzima ofanayo obulokhu udingeka ngaphambi kokuthi uthole inani lezibalo kusafanele wenziwe. Umbono kufanele ube khona, ube ngokoqobo futhi ufezeke. Umsebenzi ongcolile uwukulungiselela ukwaziswa, ukungqubuzana kwedatha nokuhlanza idatha. Lokhu kuyohlale kudingeka kwenziwe. Ekuqeqesheni i-AI, ngisho nangaphezulu. Okwamanje azikho izinqamuleli zokungenelela komuntu. Abantu kusadingeka ukuthi bachaze ama-algorithms. Kudingeka abantu babone impendulo “elungile”.

Kafushane, ukuze i-AI iphumelele, abantu badinga:

• Ukusungula ingqalasizinda. Lokhu empeleni ukumisa imingcele lapho i-AI izosebenza khona. Kumayelana nokuthi isisekelo singakwazi yini ukusekela idatha engahlelekile, i-blockchain, i-IoT, ukuphepha okufanele.
• Usizo ekutholeni. Thola futhi unqume ukutholakala kwedatha. Idatha yokuqeqesha i-AI kufanele ibe khona futhi itholakale.
• Hlela idatha. Uma kwethulwa ngesethi enkulu yedatha futhi, ngenxa yalokho, inani elikhulu lemiphumela engaba khona, kungase kudingeke isazi sesizinda ukuze sihlole imiphumela. I-Curation izophinda ihlanganise nokuqinisekiswa kokuqukethwe kwedatha.

Ukuboleka umusho ovela kososayensi bedatha, ukuze izinkampani ziphumelele nge-AI, ukuze zikwazi ukwengeza inani kumakhono okuhlaziya akhona, zidinga ukukwazi ukuhlukanisa isignali emsindweni, umlayezo ovela ku-hype.

Eminyakeni eyisikhombisa edlule, IBM's Ginni Rometty washo okuthile okufana nokuthi, I-Watson Health [AI] iyisithombe sethu senyanga. Ngamanye amazwi, i-AI - okulingana nokufika kwenyanga - iyinjongo ekhuthazayo, efinyelelekayo, enwetshiwe. Angicabangi ukuthi sesifike enyangeni. Nokho. I-IBM, nezinye izinkampani eziningi ziyaqhubeka nokusebenzela umgomo wokuguqula i-AI.

Uma i-AI kuyinyanga, inyanga iyabonakala futhi iseduze kunangaphambili.