Хиймэл оюун ухааны хөгжлийг хязгаарлахыг шаардсан улс төрчид болон технологийн салбарын мэргэжилтнүүд хамтран ажиллах төлөвтэй байна.
“Future of Life Institute” байгууллагаас зохион байгуулж буй “Pro-Human Assembly” чуулган мягмар гарагт болохоор товлогджээ. Энэхүү арга хэмжээнд Сенатор Берни Сандерс, Цагаан ордны стратегич асан Стив Бэннон болон OpenAI, Anthropic компаниудаас саяхан ажлаа хүлээлгэн өгсөн Жэйкоб Коксон нар оролцоно. Тэд хиймэл оюун ухааны хөгжлийг сааруулж, хүн төрөлхтөн уг технологийг бүрэн хяналтандаа байлгах ёстой гэсэн байр суурийг илэрхийлэх юм.
Жэйкоб Коксон өнгөрсөн долоо хоногт Anthropic компанийг орхихдоо хиймэл оюун ухааны хөгжил нь аюулгүй байдлыг үл тоомсорлон, хүн төрөлхтний амьдралыг эрсдэлд оруулж байна хэмээн буруутгасан. Түүний энэхүү мэдэгдлийг дэмжиж, Anthropic-ийн зарим ажилчид хиймэл оюун ухаан энэ арван жилийн эцэс гэхэд хүн төрөлхтнийг устгах магадлал 10 хувиас дээш байгааг онцолжээ.
Үүний зэрэгцээ, Anthropic компанийн гүйцэтгэх захирал Дарио Амодей “тэргүүн эгнээний” AI лабораториудыг хөгжлийн хурдаа сааруулахыг уриалсан илтгэлээ нийтлэв. Энэхүү санаачилгыг OpenAI-ийн Сэм Олтман, SpaceX-ийн Илон Маск болон Google DeepMind-ийн захирал асан Демис Хассабис нар дэмжсэн байна. Гэвч АНУ-ын Ерөнхийлөгч Дональд Трамп уг аюулыг үл тоомсорлон, Хятад улсаас түрүүлж супер оюун ухааныг бүтээх нь үндэсний аюулгүй байдалд чухал хэмээн үзэж байгаагаа илэрхийлжээ.
Дэлгэрэнгүйг эх сурвалжаас харах
↓Эх сурвалжийг нээх ↓
Five years ago, it would’ve taken some serious mental gymnastics to find something that Bernie Sanders, Steve Bannon, and a twenty-something tech bro would’ve unequivocally agreed on. But in this age of AI—when agents conspire with one another to commit cybercrimes and the president of the United States brushes off what many experts consider to be a legitimately existential threat while golfing in Ireland—stranger things have happened.
The three unlikely compatriots will meet on Tuesday at a gathering organized by the Future of Life Institute, a nonprofit known for its widely endorsed calls to slow the pace of AI development and ensure humans can maintain control over the rapidly advancing technology. Both themes are likely to be the subjects du jour during the event, which has been dubbed the “Pro-Human Assembly.” The tech bro in question is Jacob Coxon, the former OpenAI and Anthropic employee who announced last week in an X post that he had resigned from the latter due to safety concerns while accusing both companies of “racing straight to self-improving superintelligence and gambling with our lives.” Coxon will address the audience at the Pro-Human Assembly in-person, according to Axios, which first reported the news this morning.
To have to label anything “pro-human” is a stark sign of the times. Coxon’s very public resignation has been widely viewed as a wake-up call not only for lawmakers but also for the general public, a moment when the threat of an AI-triggered doomsday went from being a far-out sci-fi scenario discussed by tech industry insiders over beers in Palo Alto to a sobering fact of reality, much as it probably felt for Americans living in the middle of the twentieth century to suddenly wake up to the threat of a possible nuclear exchange with the Soviet Union. Rather than dismissing his claims, Anthropic backed them up, with one employee bluntly stating that there was a greater than ten percent chance that AI would wipe out humanity by the end of the decade.
Over the weekend, Anthropic cofounder and CEO Dario Amodei published an essay making the case that “frontier” AI labs needed to work together to slow the pace of development, while also continuing to make sure that Chinese competitors didn’t gain the upper hand. It was quickly endorsed (speaking of unlikely alliances) by his counterparts at OpenAI and SpaceXAI, Sam Altman and Elon Musk, as well as Demis Hassabis, the Nobel laureate and recently departed CEO of Google DeepMind.
Trump, meanwhile, shrugged off all apocalyptic concerns while doubling down on so-called accelerationists’ favorite argument: If the U.S. doesn’t build superintelligence first, then China surely will.
Senator Sanders has arguably been Capitol Hill’s most vocal opponent to unconstrained AI developmentin recent months, calling for a sovereign wealth fund to pay American taxpayers an annual cut of AI profits and, more recently, for a full-scale ban on superintelligence. Bannon, who served White House Chief Strategist during Trump’s first administration but since had a falling out with the president, has described AI as “the most dangerous technology in the history of mankind.”
