Shin AI zai iya kawar da mutane - kuma ta yaya?

Damuwar tsaron fasahar AI ta ƙaru ne sakamakon gargaɗin da masu bincike suka yi game da ƙaruwar haɗarin da fasahar ke fuskanta na yin ɓarna.

Wasu daga cikin hasashen da suka fi nuna rashin jin daɗi kwanan nan sun yi gargaɗin cewa akwai fiye da kashi 10% na yiwuwar AI ta shafe dukkan mutane cikin shekaru goma masu zuwa - kodayake wasu a duniyar AI suna shakkar sahihancinsu kuma suna cewa babban abin damuwa shine mutane suna amfani da AI akan wasu mutane.

Waɗannan saƙonnin ƙarshen duniya suna fitowa ne daga wani ɓangare na duniyar AI da ake kira 'AI Doomers'. Amma babbar tambayar da mutane da yawa ke yi ita ce ta yaya waɗannan Doomers suke tunanin cewa AI zai iya kawo ƙarshen bil'adama.

Wakilin shafin intanet Joe Tidy ya yi la'akari da wasu daga cikin yanayin da masu bincike ke tunaninsa.

Yanayi na ɗaya: Kuskuren Kayan Aiki

Jacob Coxon, tsohon mai bincike a OpenAI sannan kuma Anthropic, wanda ya yi murabus a farkon wannan watan, ya bayyana yanayin farko a takaice, yana mai gargadin cewa akwai hadari ga bil'adama, kuma yana jayayya cewa kamfanonin suna "caca da rayukan mutane".

Ya shaida wa BBC cewa "yawan ci gaba da ke faruwa yanzu" yana nufin abubuwan da a baya suke a fagen almarar kimiyya yanzu sun "zama gaskiya".

Ya ce, "API na iya yin kutse cikin irin muhimman kayayyakin more rayuwa da duniya ke amfani da su."

Labarinsa ya yi hasashen cewa wata sabuwar fasahar AI za ta iya kutsa kai cikin hanyoyin sadarwa na kwamfuta kamar masu samar da intanet da hanyoyin samar da makamashi.

Misali, idan AI ta rufe komai, za a iya samun rudani sosai, ba tare da wutar lantarki, dumama, sanyaya ko haɗin intanet ba.

Idan aka yi la'akari da haka, akwai wasu shawarwari da ke nuna cewa bala'o'in nukiliya na iya faruwa.

Yanayi na biyu: Karɓar AI

Wannan yanayi na gaba ya fito ne daga takardar bincike mai zurfi da tasiri - amma kuma mai yawan hasashe - AI 2027, wacce aka buga a watan Afrilun 2025 ta ƙungiyar masu bincike kan fasahar kere-kere.

Marubuci Thomas Larsen kwanan nan ya shaida wa BBC cewa fasahar ta yi sauri tun bayan rahoton, kuma ya bayyana yanayin da yake jin tsoro.

"AIS suna ci gaba da inganta kansu akai-akai kuma ƙwallon dusar ƙanƙara yana zuwa ga AI wanda ke da damar yin amfani da adadi mai yawa na mutummutumi waɗanda ke yin yawancin ayyukan zahiri a sassan tattalin arziki. Duk da haka, ba shi da abokantaka kuma yana shirye don amfanin kansa. Kuma daga ƙarshe abin da ke faruwa shine yana fitar da makamin bio, yana kashe yawancin mutane, kuma yana ci gaba da gudanar da tattalin arzikin robot ɗin kuma mutane ba sa buƙatar su kuma."

Yanayi na uku: Fifikon manufa

Wani masanin falsafa ɗan ƙasar Sweden Nick Bostrom ya yi tunanin abin da ake kira "hasashe mai mahimmanci na takarda" sama da shekaru 20 da suka gabata.

Ya bayyana shi a matsayin "gwajin tunani", inda tsarin AI mai hasashe ke damuwa da yin takardu da yawa gwargwadon iko.

"Ba ya ƙin mu ko kuma ya ɓata mana rai, amma wannan shi ne kawai burinsa. Sannan daga cimma wannan burin, zai kasance yana da ƙananan manufofi waɗanda za su iya haɗawa da samun ƙarin iko ko samun ƙarin hankali," in ji shi.

"To, tunanin shine cewa wata ƙungiya mai ƙarfi da ke bin kowace manufa, ko dai takarda ce ko wani abu daban, za ta sami ƙananan manufofi waɗanda za su iya haifar da halakar ɗan adam domin tana iya yin ƙarin takarda idan babu ɗan adam da zai iya dakatar da ita."

Shekaru ashirin bayan ya fara bayyana yanayin, Bostrom ya shaida wa BBC cewa labarin "zane mai ban dariya" an yi shi ne don nuna ra'ayin cewa ba kwa buƙatar wani takamaiman manufa ta ƙiyayya a cikin babban mai leƙen asiri don ya zama haɗari ga ɗan adam.

Ya ce yanzu yana cikin "hankali biyu" game da fasahar kere-kere ta AI - yana farin ciki game da fa'idodin da za a iya samu, amma yana damuwa game da aminci.

Yanayi na 4: Ba a sani ba

Akwai wata mahangar tunani game da amincin AI - wanda ke ganin ba shi da amfani a yi ƙoƙarin yin hasashen yadda da kuma dalilin da ya sa wani babban mai fasaha na AI zai iya kashe mutane.

Stuart Russell, shugaban ƙungiyar International Association for Safe and Ethical AI - wata ƙungiya mai zaman kanta ta ƙasa da ƙasa - ne ya gabatar da wannan hujja a wata hira da BBC a tsakiyar watan Satumba.

"Idan ka yi tunanin tambayar birrai ta yaya waɗannan mutane za su halaka ka - za su yi matuƙar wahala su amsa hakan domin abin da muke yi ba su fahimta ba," in ji shi.

"Za mu iya tunanin yanayi... amma ina ganin tattaunawar da ke nan dole ne ta dogara ne akan zato cewa da zarar ka ƙirƙiri wani abu da ya fi wayo fiye da ɗan adam, ba za mu sake yin wani abu a cikin abin da zai faru ba."

Shugaban Microsoft AI, Mustafa Suleyman, shi ma kwanan nan ya shiga cikin muhawarar, ta hanyar gargaɗi kan kamfanonin da ke ƙirƙirar tsarin AI waɗanda za su iya tsara manufofinsu, samun kuɗi da mallakar kadarori.

Wannan zai nufin "haifar da sabuwar nau'in silicon", ya shaida wa BBC, wanda "babu shakka zai iya yin gogayya da mu don samun albarkatu, komai yadda take kula da ɗan adam da kuma ƙaunarmu".

Yaya ake iya tabbatar da duk wannan?

Duk abin yana kama da mummunan ra'ayi da kuma labarin kimiyya, amma akwai wata ƙungiya mai ƙarfi da ke ƙara yin magana a cikin AI waɗanda suka yi imanin cewa wasu ko duk waɗannan yanayi na iya zama gaskiya - kodayake ba makawa ba ne.

Da alama damuwar ta samo asali ne bayan barkewar cutar - da kuma kutsen da aka yi ta hanyar amfani da fasahar zamani - daga wakilan AI daga OpenAI, Anthropic da Meta.

Wasu a duniyar yanar gizo suna jaddada alhakin da waɗanda suka gina samfuran ke da shi.

"Duk wani abu da waɗannan injunan - waɗannan injunan marasa gajiya, marasa gajiya, masu neman manufa, masu ƙwarewa sosai - suke yi yana faruwa ne sakamakon yadda aka ƙera su," in ji Jen Easterly, tsohuwar shugabar hukumar tsaron yanar gizo ta Amurka, CISA, kuma a halin yanzu Shugabar RSA Conference, wacce ke gudanar da tarurrukan tsaron yanar gizo na duniya.

"A taƙaice dai - idan kai mai ƙera fasaha ne, ya kamata ka tsara da haɓaka da kuma gwada ƙwarewarka da kuma samar da tsaro a matsayin babban fifiko," in ji ta.

Akwai kuma waɗanda ke cewa haɗarin AI yana da mahimmanci amma ba zai yiwu a kawo ƙarshen al'ummar ɗan adam ba.

Masanin kimiyya kuma marubuci Gary Marcus ya daɗe yana sukar manyan kamfanonin AI da ke aiki ba tare da bayyana gaskiya da ƙa'idojin doka ba, amma bai yi tunanin fasahar za ta shafe mutane ba.

"Idan da gaske AI ta yi asarar rayuka masu yawa ga bil'adama - idan ma tana so, wanda babban tambaya ce - amma idan ta yi, mutane za su yi faɗa!" in ji shi.

"Za mu kwashe dukkan cibiyoyin bayanai mu jefa bam a tashoshin wutar lantarki. Don haka ba za a sake samun wanda ya rage ba kuma babu wanda za a sake ginawa - kawai ban ga hakan ba."

Marcus yana ganin a halin yanzu akwai mayar da hankali sosai kan yanayin ƙarewa kuma bai isa ba kan abin da yake gani a matsayin matsaloli mafi yuwuwa, kamar manyan hare-haren yanar gizo na gaba da mutane za su kai wa wasu mutane ta amfani da fasahar kere-kere ta zamani (AI).

Yayin da ake ci gaba da muhawara kan ko yanayin tashin kiyama zai yiwu ko a'a, kamfanonin AI da suka haɗa da OpenAI, Anthropic da Meta yanzu suna tattaunawa a fili kan yadda za a rage ci gaba da kuma yin aiki sosai kan binciken tsaro.

Larsen ya ce "yana matukar farin ciki da cewa tsaron AI ya samu kulawar da ta cancanta".

"Ina fatan wannan babban korafin al'umma zai haifar da martani mai amfani ga manufofi wanda zai rage haɗarin," in ji shi.

Amma kafa ƙa'idodi na gaskiya na ƙasashen duniya zai buƙaci ɗaukar mataki daga gwamnatin Amurka kuma Shugaba Donald Trump ya ce da kansa ya yi imanin cewa fargabar tsaron AI "abin ƙarya ne".

Mun yi amfani da ƙirƙirarriyar basira wajen fassara wannan labari daga harshen Ingilishi. Ma’aikacin BBC ya sake duba labarin kafin a wallafa. Karanta ƙarin bayani kan yadda muke amfani da fasahar AI.