Njengoko kuvela, ewe, kodwa nje kancinci

I-AI ikho yonke indawo. Enye yezona ndawo zixhaphakileyo ze-AI ekhaya kwezi ntsuku yi-smartphone, amakhaya ahlakaniphile kunye nezixhobo zombane. Kutshanje, njengoko sasihleli phantsi kwisidlo sangokuhlwa, saba nencoko kunye ne-Alexa eyahamba ngolu hlobo:

Me: I-Alexa, dlala amanqaku aphezulu e-Cubs. [Eli lilungu eliboniswe kwiscreen sasekhaya se-Alexa. Cela i-Alexa idlalele iqela lakho olithandayo.]

Alexa: Ndifumene into kwiwebhu. [Ndiyazi xa i-Alexa iqala ngolu hlobo, kukho ingxaki. Ayizuhamba kakuhle. I-Alexa ibonisa uluhlu lweevidiyo ezimbalwa. Kuyavunywa, uninzi lwazo ziividiyo zebaseball ezinabadlali abenza imidlalo ekhethekileyo kule minyaka mihlanu idlulileyo. Ngenxa yam. Zama kwakhona.]

Me: Alexa, ndibonise amagqabantshintshi omdlalo webaseball wamva nje weChicago Cubs. [Ndiyathemba ukuba ayizukubona ithoni ezithobayo njengoko ndisamkela ityala lokungakwazi ukusiqonda isicelo sam.]

Alexa: Iimbalasane zomdlalo weChicago Cubs ziya kufumaneka kwiiyure ezimbini emva kokuba umdlalo uphelile. [Inkqubela. Ndandingazi ukuba inokuba bayadlala ngawo lo mzuzu. Ithamsanqa lisecaleni lam. Ngephanyazo ndinethemba.]

Me: Alexa, kulungile ndibonise iimbalasane zayizolo, ke. [Ewe, ukudana kwam kuyaqala ukubonakala. Ndisondele kakhulu ekwaphuleni ikhowudi. Ndiphantse ndayingcamla.]

Alexa: Uxolo andiyazi lonto. [Ithetha le nto rhoqo kakhulu. Mhlawumbi bendingazicacelanga.]

Mna: Ingaba uyadlala ngam? Dlala, iimbalasane zevidiyo zomdlalo weMajor League baseball phakathi kweChicago Cub kunye nePittsburgh Pirates ngoMvulo wama-25 kaJulayi 2022 eWrigley Field. [Ngeli xesha ndiqinisekile ukuba ndiyibethelele. Nditshice isicelo esicacileyo, esingacacanga eso sisakhono endisaziyo ukuba i-Alexa inaso. Ikwenzile oku ngaphambili. ]

alexa: [Kwathi cwaka. Akukho nto. Akukho mpendulo. Ndilibele ukuthetha igama lomlingo lokuvusa, Alexa.]

The I-IQ ephakathi oneminyaka eli-18 ubudala ujikeleze 100. I-IQ eqhelekileyo yomntu oneminyaka emi-6 ubudala ngama-55. I-Google AI IQ yavavanywa ukuba ibe ngu-47. I-IQ kaSiri iqikelelwa ukuba yi-24. I-Bing ne-Baidu ikwi-30. Khange ndilufumane uvavanyo lwe-Alexa's IQ, kodwa amava am ayefana nokuthetha nomntwana ongekayo esikolweni.

Abanye banokuthi, akulunganga ukunika ikhompyuter uvavanyo lwe-IQ. Kodwa, le ngongoma ngokugqibeleleyo. Isithembiso se-AI kukwenza okwenziwa ngabantu, ngcono kuphela. Ukuza kuthi ga ngoku, intloko-to-intloko nganye-okanye, siya kuthi, inethiwekhi ye-neural kwinethiwekhi ye-neural-umngeni ugxininise kakhulu. Ukudlala ichess. Ukuxilonga isifo. Ukusenga iinkomo. Ukuqhuba iimoto. Irobhothi idla ngokuphumelela. Into endifuna ukuyibona nguWatson esenga inkomo eqhuba imoto edlala iJeopardy. Ngoku, ukuba iya kuba yi-trifecta. Abantu abakwazi nokukhangela imidiza yabo ngelixa beqhuba ngaphandle kokufumana ingozi.

I-IQ ye-AI

Ifakwe ngumatshini. Ndiyakrokra ukuba andindedwa. Ndaye ndacinga, ukuba le yimeko yobugcisa, zihlakaniphe kangakanani ezi zinto? Ngaba sinokuthelekisa ubukrelekrele bomntu nomatshini?

Izazinzulu ziyahlola ubuchule bokufunda nokuqiqa. Ukuza kuthi ga ngoku, abantu bokwenziwa abenzanga into eyinyani. Abaphandi basebenzisa ukusilela ukuchonga izikhewu ukuze siqonde ngcono apho kufuneka kwenziwe uphuhliso olongezelelweyo kunye nenkqubela phambili.

Ukuze ungaphoswa yingongoma kwaye ulibale ukuba "mna" kwi-AI imele ntoni, abathengisi ngoku baqulunqe igama elithi Smart AI.

Ngaba i-AI inomdla?

Ngaba iirobhothi zineemvakalelo? Ngaba iikhompyutha amava emotions? Hayi masiqhubeke. Ukuba uyafuna ukufunda ngayo, enye (yangaphambili) injini kaGoogle ibango imodeli ye-AI uGoogle asebenza kuyo inengqondo. Wayenencoko eyothusayo kunye ne-bot eyamenza waqiniseka ukuba ikhomputha ineemvakalelo. Ikhompyuter yoyikela ubomi bayo. Andikholelwa nokuba ndisibhale eso sivakalisi. Iikhompyutha azinaloyiko lobomi. Iikhompyutha azikwazi ukucinga. Ii-algorithms azicingwa.

Noko ke, andiyi kumangaliswa, nangona kunjalo, ukuba ikhompyutha iphendula kumyalelo kungekudala ngokuthi: “Uxolo, Dave, andinakuyenza loo nto.”

Iyasilela phi i-AI?

Okanye, ngokuchanekileyo, kutheni iiprojekthi ze-AI zisilela? Bayasilela ngenxa yezizathu ezifanayo ukuba iiprojekthi ze-IT zihlala zisilela. Iiprojekthi ziyasilela ngenxa yolawulo olugwenxa, okanye ukusilela ekulawuleni ixesha, ububanzi okanye uhlahlo lwabiwo-mali..:

• Umbono ongacacanga okanye ongachazwanga. Isicwangciso esibi. Usenokuba uve abaphathi besithi, "Sifuna nje ukujonga ibhokisi." Ukuba isiphakamiso sexabiso asikwazi ukuchazwa, injongo ayicacanga.
• Ulindelo olungekho ngqiqweni. Oku kusenokubangelwa kukungaqondani, unxibelelwano oluhle, okanye ucwangciso olungekho ngqiqweni. Ulindelo olungeyonyani lusenokubangelwa kukunqongophala kokuqonda kwezakhono zezixhobo ze-AI kunye nendlela yokusebenza.
• Iimfuno ezingamkelekanga. Iimfuno zeshishini azichazwanga kakuhle. Iimetriki zempumelelo azicacanga. Kwakhona kolu didi kukungaxabiseki kwabasebenzi abayiqondayo idatha.
• Iiprojekthi ezingaqingqwanga nezingacingelwayo. Iindleko azikaqikelelwa ngokupheleleyo nangokungakhethi cala. Iimeko ezinokuvela azikacwangciswa kwaye azicingelwanga. Igalelo lexesha labasebenzi abasele bexakeke kakhulu lijongelwe phantsi.
• Iimeko ebezingalindelekanga. Ewe, ithuba liyenzeka, kodwa ndicinga ukuba oku kuwela phantsi kocwangciso olubi.

Jonga, kwakhona, iposti yethu yangaphambili Izizathu ze-12 zokungaphumeleli kwi-Analytics kunye noBuchule boShishino.

I-AI, namhlanje, inamandla kakhulu kwaye inokunceda iinkampani ukuba zifezekise impumelelo enkulu. Xa amaphulo e-AI engaphumeleli, ukusilela kungasoloko kulandelelwa kwenye yezi zingentla.

Ingaba iphi i-AI Excel?

I-AI ilungile kwimisebenzi ephindaphindiweyo, enzima. (Ukuba nobulungisa, inokwenza imisebenzi elula, engaphindaphindiyo, nayo. Kodwa, kuya kuba nexabiso eliphantsi ukwenza umntwana wakho ongekabinasikolo ayenze loo nto.) Kuhle ekufumaneni iipateni kunye nobudlelwane, ukuba bukhona, kwiimali ezininzi zedatha.

• I-AI yenza kakuhle xa ikhangela imisitho engahambelani neepateni ezithile.
  • Ukufumanisa ubuqhetseba bekhadi letyala imalunga nokufumana iitransekshini ezingalandeli iipatheni zosetyenziso. Ithanda ukwenza impazamo kwicala lokulumkisa. Ndifumene iifowuni ezisuka kwikhadi lam letyala nge-algorithm enenzondelelo egqithisileyo xa ndandigcwalise imoto yam yokurenta ngegesi eDallas ndaza ndazalisa imoto yam eChicago. Yayisemthethweni, kodwa ingaqhelekanga ngokwaneleyo ukuba ifakwe iflegi.

"American Express iinkqubo $ 1 yetriliyoni in transactions kwaye 110 million amakhadi AmEx esebenzayo. Bathembele kakhulu kuhlalutyo lwedatha kunye ne-algorithms yokufunda koomatshini ukunceda ukufumanisa ubuqhophololo ngexesha lokwenyani, ngoko ke konga izigidi kwilahleko”.

• Ubuqhophololo kunye nokuxhatshazwa kwamachiza. Iinkqubo zinokufumana iipateni ezingaqhelekanga zokuziphatha ngokusekelwe kwimithetho emininzi ecwangcisiweyo. Ngokomzekelo, ukuba isigulane sabona oogqirha abathathu abahlukeneyo edolophini ngosuku olunye kunye nezikhalazo ezifanayo zentlungu, uphando olongezelelweyo lunokuqinisekiswa ukuba lukhuphe ukusetyenziswa kakubi.
• AI ngaphakathi ukhathalelo lwempilo uye waba nempumelelo ebalaseleyo.
  • I-AI kunye nokufunda okunzulu kwafundiswa ukuthelekisa i-X-rays kunye neziphumo eziqhelekileyo. Ikwazile ukunyusa umsebenzi weengcali zeradiology ngokuxela izinto ezingaqhelekanga ukuba ingcali yeradiologist ijongiwe.
• I-AI isebenza kakuhle nayo ezentlalo kunye nokuthenga. Esinye isizathu sokuba sibona oku kakhulu kukuba kukho umngcipheko ophantsi. Umngcipheko wokuba i-AI ingalunganga kwaye ibe neziphumo ezibi kakhulu iphantsi.
• • Ukuba uthandile/uthengile lo, sicinga ukuba uya kuthanda oku. Ukusuka kwiAmazon ukuya kwiNetflix kunye neYouTube, bonke basebenzisa uhlobo oluthile lokuqatshelwa kwepateni. I-Instagram AI ijonga unxibelelwano lwakho ukuze ugxile kwisondlo sakho. Oku kuthanda ukusebenza ngcono ukuba i-algorithm inokubeka izinto ozikhethayo kwibhakethi okanye kwiqela labanye abasebenzisi abenze ukhetho olufanayo, okanye ukuba umdla wakho umxinwa.
  • I-AI iyonwabele impumelelo ethile ukuqaphela ubuso. I-Facebook iyakwazi ukuchonga umntu ofakwe ngaphambili kwifoto entsha. Ezinye iinkqubo ezinxulumene nokhuseleko zakwangoko zikhohliswe ziimaski.
• I-AI iyonwabele impumelelo zolimo usebenzisa ukufunda koomatshini, iinzwa ze-IoT kunye neenkqubo ezixhunyiwe.
  • I-AI incedise iitrektara ezikrelekrele amasimi okulima kunye nokuvuna ukwandisa isivuno, ukunciphisa isichumiso kunye nokuphucula iindleko zemveliso yokutya.
  • Ngamanqaku edatha evela kwiimephu ze-3-D, izinzwa zomhlaba, iidrone, iipatheni zemozulu, esweni yokufunda umatshini ifumana iipateni kwiiseti zedatha enkulu ukuqikelela elona xesha lilungileyo lokutyala izityalo kunye nokuqikelela izivuno phambi kokuba zityalwe.
  • Iifama zobisi sebenzisa iirobhothi ze-AI ukuba iinkomo zizise ngokwazo, i-AI kunye nokufunda koomatshini kwakhona ukubeka iliso kwiimpawu ezibalulekileyo zenkomo, umsebenzi, ukutya kunye nokuthatha amanzi ukuze bahlale besempilweni kwaye banelisekile.
  • Ngoncedo lwe-AI, mafama abangaphantsi kwe-2% yabemi bondla izigidi ezingama-300 kulo lonke elase-USA.
  • Ubukrelekrele bokwenziwa kwezoLimo

Kukho amabali amnandi e-AI impumelelo kumashishini eenkonzo, kwiivenkile, kumajelo eendaba nakwimveliso. I-AI ngokwenene yonke indawo.

Amandla e-AI kunye nobuthathaka obuthelekisiweyo

Ukuqonda okuqinileyo amandla kunye nobuthathaka be-AI kunokuba negalelo kwimpumelelo yamanyathelo akho e-AI. Khumbula, kwakhona, ukuba izakhono ngoku kwikholamu yasekunene ngamathuba. Ezi ziindawo apho abathengisi kunye nabamkeli be-blood edge benza inkqubela ngoku. Siza kujonga izakhono ezicela umngeni kwi-AI ngoku kwakhona enyakeni kwaye sibhale ishifti yasekhohlo. Ukuba ufunda le tshathi ilandelayo ngononophelo, andiyi kumangaliswa ukuba kukho intshukumo ethile phakathi kwexesha endibhala ngalo kunye nexesha elipapashwa ngalo.

Lithini ikamva le-AI?

Ukuba i-AI ibikrelekrele, inokuqikelela ukuba ikamva liphethe ntoni. Kucacile ukuba zininzi iingcamango eziphosakeleyo malunga nento i-AI enokuyenza kwaye ayikwazi ukuyenza. Abaninzi iingcamango ezigwenxa kunye nokungafundi kwe-AI ziziphumo zentengiso yetekhnoloji ngaphezulu-hyping amandla akhoyo. I-AI iyamangalisa into enokuyenza namhlanje. Ndixela kwangaphambili ukuba ubuthathaka obuninzi kwikholamu yesandla sokunene buya kutshintshela ngakwesobunxele kwaye bube namandla kwi-2 okanye kwi-3 iminyaka ezayo.

[Emva kokuba ndiligqibile eli nqaku, ndiye ndabonisa isiqendu esingaphambili I-OpenAI, i-generator yolwimi lweqonga le-AI evulekileyo. Usenokuba ubone ubugcisa obuveliswe yi-DALL-E yayo. Ndandifuna ukwazi ukuba icinga ntoni ngekamva le-AI. Nantsi into ebeyithetha. ]

Ikamva le-AI alikho malunga nokuthenga iiseva ezimbalwa kunye nokufaka iphakheji yesoftware engaphandle kweshelufu. Imalunga nokufumana kwaye uqeshe abantu abafanelekileyo, ukwakha iqela elichanekileyo, kunye nokwenza utyalo-mali olufanelekileyo kuzo zombini izixhobo kunye nesoftware.

Ezinye iimpumelelo ezinokubakho ze-AI kule minyaka imbalwa izayo ziquka:

• Ukwandisa ukuchaneka kweengqikelelo kunye neengcebiso
• Ukuphucula iinkqubo zokwenza izigqibo
• Ukukhawulezisa uphando kunye nophuhliso
• Ukunceda ukuzenzekelayo kunye nokwandisa iinkqubo zoshishino

Nangona kunjalo, kukwakho ukusilela okunokwenzeka kwe-AI ekufuneka amashishini ayazi, njengale:

• Ukuthembela ngokugqithisileyo kwi-AI ekhokelela kwizigqibo ezisezantsi
• Ukungaqondi indlela i-AI esebenza ngayo ekhokelela ekusebenziseni kakubi
• I-bias kwidatha esetyenziselwa ukuqeqesha iimodeli ze-AI ezikhokelela kwiziphumo ezingachanekanga
• Ukhuseleko kunye neenkxalabo zabucala malunga nedatha esetyenziselwa ukuqeqesha iimodeli ze-AI

Ke, kuthetha ukuthini oku kumashishini atyala imali kwi-AI ukongeza kuhlahlelo lwabo lwemveli? Impendulo emfutshane kukuba, akukho ndlela-mfutshane. I-85% yamalinge e-AI ayiphumeleli. Okubangela umdla kukuba, oku kufana neenkcukacha-manani ezicatshulwe rhoqo ezinxulumene neeprojekthi ze-IT zemveli kunye ne-BI. Umsebenzi onzima ofanayo osoloko ufunwa ngaphambi kokuba ufumane ixabiso kuhlalutyo kufuneka usenziwa. Umbono kufuneka ubekho, ube yinyani kwaye ube nokufezekiswa. Umsebenzi ongcolileyo kukulungiswa kwedatha, ukuphikisana kwedatha kunye nokucoca idatha. Oku kuya kusoloko kufuneka kwenziwe. Ekuqeqesheni i-AI, nangakumbi. Okwangoku azikho iindlela ezimfutshane zongenelelo lwabantu. Abantu kusafuneka bachaze ii-algorithms. Abantu kufuneka bachonge impendulo "echanekileyo".

Ngamafutshane, ukuze i-AI iphumelele, abantu kufuneka:

• Ukuseka iziseko zophuhliso. Oku kukuseka imida apho i-AI iza kusebenza kuyo. Ku malunga nokuba isiseko sinokuxhasa idatha engacwangciswanga, i-blockchain, i-IoT, ukhuseleko olufanelekileyo.
• Uncedo ekufumaneni. Fumana kwaye uqikelele ukufumaneka kwedatha. Idatha yokuqeqesha i-AI kufuneka ibekho kwaye ifumaneke.
• Lungelelanisa idatha. Xa kuboniswa ngedatha enkulu kwaye, ngenxa yoko, inani elikhulu leziphumo ezinokuthi zifumaneke, ingcali yesizinda inokufuneka ukuba ivavanye iziphumo. I-Curation iya kubandakanya ukuqinisekiswa komxholo wedatha.

Ukuboleka ibinzana kwizazinzulu zedatha, ukuba iinkampani ziphumelele nge-AI, ukuze zikwazi ukongeza ixabiso kwizinto ezikhoyo zokuhlalutya, kufuneka zikwazi ukwahlula isignali kwingxolo, umyalezo ovela kwi-hype.

Kwiminyaka esixhenxe eyadlulayo, IBM's Ginni Rometty uthe into enje, iWatson Health [AI] yinyanga yethu. Ngamanye amazwi, i-AI - elilingana nokufika kwenyanga - yinjongo ekhuthazayo, enokufezekiswa, eyolula. Andiqondi ukuba siye sahlala enyangeni. Ukanti. I-IBM, kunye nezinye iinkampani ezininzi ziyaqhubeka nokusebenza ngenjongo yokuguqula i-AI.

Ukuba i-AI yinyanga, inyanga iyabonakala kwaye isondele ngakumbi kunokuba ibikhona.