Shaqaalaha AI-da ayaa mustaqbalka aadanaha u baqaya sida uu BBC-da u sheegay mid horay uga shaqayn jiray

Cilmi-baare ku takhasusay Garaadka Macmalka (AI) kaas oo iska casilay shirkadda Anthropic, ayaa sheegay in dadka ka shaqeeya tiknoolajiyaddan ay "si dhab ah uga baqayaan" sida dheereysa ay u horumareyso iyo waxa ay aadanaha uga dhigan tahay.

Jacob Coxon ayaa Sabtidii wariyaha BBC-da Laura Kuenssberg uga warramay qoraalkiisii is-casilaadda ee sida weyn baraha bulshada ugu faafay, kaas oo uu ku muujiyay walaac ku saabsan AI-da aan la xakamayn karin.

"Waxaan aaminsannahay haddii aynaan gaabin xawaaraha ay hadda AI-da ku socoto, waxaa dhici karta in dhammaanteen aan mustaqbalka dhow dhiman karno," ayuu yiri.

Is-casilaaddiisa waxay ku soo beegantay xilli walaaca laga qabo AI-da uu ku sii kordhayo. Madaxa shirkadad Anthropic Dario Amodei ayaa maqaal uu qoray kaga hadlay in loo baahan yahay in la dhimo xawaaraha AI-da.

Madaxda laba shirkadood oo AI ah oo la tartama Anthropic, Sam Altman oo ka tirsan OpenAI iyo Elon Musk oo ka tirsan xAI, waxay labadooduba sheegeen in ay taageersan yihiin soo jeedinta Amodei ee ah in guud ahaan la dhimo xawaaraha ee AI-da ku socoto, lana sameeyo xeerar lagu nidaamiyo, iyo sidoo kale in si madax-bannaan loo kormeero horumarinta noocyada AI.

Amodei wuxuu sheegay in horumarinta tiknolojiyaddan aan su'aal la gelinayn, balse khataraha la xiriira ay yihiin kuwo "halis ah", isla markaana shirkadaha iyo dowladuhu u baahan yihiin waqti ay khatarahaas wax uga qabtaan.

Coxon oo 27 jir ah, una soo shaqeeyay OpenAI ka hor inta uusan ku biirin Anthropic, ayaa soo dhaweeyay fikradda ah in horumarka la gaabiyo, balse wuxuu sheegay in arrintaasi u baahan tahay in Shiinuhu qeyb ka noqdo si looga fogaado "tartan heer caalami ah".

"Dadka ka shaqeeya shirkadahan aad bay dhab uga tahay marka ay dalbanayaan in la sameeyo xeerar, sababtoo ah waxay isku arkaan in tartanka uu qasbayo in ay sii horumariyaan AI-da. Waxayna ka baqayaan natiijada tartankaas," ayuu u sheegay BBC-da.

Sida uu sheegay Coxon, su'aasha ugu adag ee laga jawaabi karo waxay ahayd waxa ay u ekaan lahayd masiibo AI oo dunida ku habsata.

Mid ka mid ah khataraha uu Amodei tilmaamay ayaa ahaa in amarro AI ah ay u dhaqmi karaan sida kombiyuutar aad u awood badan, kuwaas oo la wareegi kara internet-ka.

Coxon wuxuu sheegay in xaaladdan ay dhab ahaan dhici karto isla sanadkan muddo lix bilood gudaheed ah.

Afhayeen u hadlay Anthropic oo ka jawaabayay shaqo ka tagista Coxon ayaa BBC-da u sheegay, "Mar walba waxaan si cad u sheegnaa in AI-da ay keeni doonto faa'iidooyin aad u waaweyn iyo khataro aan hore loo arag.

"Si aan khatarahaas wax uga qabanno, waxaan sii wadeynaa in aan dhisno noocyo leh qaar ka mid ah hababka ugu xooggan ee badbaadada."

Qaar ka mid ah dadka ku howlan AI-da ayaa sheegay in ay suuragal tahay in hadallada ku saabsan khataraha iyo awoodda AI-da loo adeegsado abuurista buunbuunin. Dhaliilayaal kale ayaa sheegaya in Anthropic ay isku dayaysay in ay kiciso dadaal sharciyeed oo lagu xannibayo tartanka, taasoo Anthropic iyo OpenAI ka dhigi lahayd laba shirkadood oo suuqa si weyn u wada maamula.

Waxaa la sheegayaa in shirakdda Anthropic ay isu diyaarineyso iibinta saamiyo dadka ay iibsadaan

Aragtida Coxon waxaa sidoo kale su'aal geliyay madaxa shirkadda AI-da ee Hugging Face, Clement Delangue.

"Waan ka xumahay, Jacob in wax laga weydiiyo khatarta AI-da waxay la mid tahay in ninka kuu dayactiro qaboojiyaha gurigaaga aad wax ka weydiiso isbeddelka cimilada," ayuu ku qoray X. "Ma dhahayo arrintu khasab ahaan ma xiiso badna ama waa khalad, laakiin aan wax walba ku eegno miisaankooda saxda ah."

Madaxa Nvidia, Jensen Huang, ayaa sidoo kale kaga hadlay sheekada Coxon shir uu toddobaadkii hore u qabtay bangiga maalgashiga Goldman Sachs. Dad dhowr ah oo shirka joogay ayaa BBC u sheegay in Huang uu hadallada Coxon ku tilmaamay kuwo aan run ahayn.

Coxon wuxuu sheegay in saaxiibbadiis iyo dadka ay wada shaqeeyaan ay ka baqayaan in khatartu dhacdo labada sano ee soo socda.

Wuxuu sheegay in shaqaalaha shirkadaha AI-da ay "qorsheynayaan waxa ay noloshooda ku samayn doonaan, isla markaana ka fikirayaan saameynta shaqadoodu leedahay", halka qaar kalena ay "ka fikirayaan in ay meel ka iibsadaan dhul, sababtoo ah waxay aad uga baqayaan xasillooni-darrada ka dhalan karta horumarka degdegga ah ee AI-da".

Si kastaba ha ahaatee, Coxon wuxuu sidoo kale qabaa rajo wanaagsan oo ku saabsan mustaqbalka AI-da. Wuxuu BBC u sheegay in dadka baaraya tiknolojiyaddan ay "si dhab ah u doonayaan in ay arkaan dhinaca wanaagsan" sida waxyaabo ay ka mid yihiin "xallinta cudurrada iyo horumarinta nolosha qof walba".

Dad caan ah oo hormuud ka ah cilmiga AI-da ayaa sanado badan ka digayay khataraha badbaadada ee tiknolojiyaddu keeni karto. Madaxda OpenAI, Google DeepMind iyo Anthropic ayaa digniino noocaas ah soo jeediyay sanadkii 2023-kii.

Digniinahan ayaa toddobaadyadii u dambeeyay sii xoogaystay, iyadoo noocyo cusub ay muujinayaan awoodo sii kordhaya oo aad u sarreeya, kuwaas oo laga yaabo in ay ku adkaato abuurayaashooda in ay xakameeyaan.

Anthropic waxay bishii April sheegtay in ay ka joojisay in dadweynuhu isticmaalaan nooca Mythos, sababtoo ah wuxuu si madax-bannaan uga baxsan karaa amarada la siiyay.