Meta chief Mark Zuckerberg talks up AI on par with human intelligence, promises to make ‘widely available’ to public
Mark Zuckerberg, chief executive of Facebook and Instagram parent Meta Platforms, has been criticised for saying he intends for the company to build artificial general intelligence (AI) systems and make them “widely available” to the public.
An industry observer called the AI plans “vaporous”, while a member of the UN’s AI body said the idea of making such powerful tools widely available was “irresponsible”.
“Our long-term vision is to build general intelligence, open source it responsibly, and make it widely available so everyone can benefit,” Zuckerberg said in a post on Instagram.
He said the company was bringing the operations of two of its AI research groups, Fundamental AI Research (FAIR) and its generative AI (GenAI) group, “closer together” and is training its next-generation Llama 3 language model.
He added that the company is building one of the largest stockpiles of Nvidia GPUs in the world to support its plans, with about 350,000 top-end Nvidia H100 chips and the processing equivalent of roughly 600,000 H100s including orders of less-powerful A100 processors, by the end of this year.
External research has estimated Meta’s 2023 orders of H100 chips at 150,000, comparable to those ordered by Microsoft and about three times larger than the next-largest orders.
Zuckerberg’s remarks late last week represented a shift for Meta, which in recent years has emphasised its plans for the “metaverse”, referring to immersive social environments that depend on the use of a headset, to the extent of changing the corporation’s name.
The goal of building general artificial intelligence refers to the nebulous concept of an AI with thinking capabilities comparable to those of a human.
“Honestly, the ‘general intelligence’ bit is just as vaporous as the ‘metaverse’,” David Thiel, a data architect and chief scientist of the Stanford Internet Observatory, told Rolling Stone magazine.
Wendy Hall, a professor of computer science at the University of Southampton and member of the UN’s advisory body on AI, said that such systems are “many years away” in practice but that even so it was “so irresponsible” of a company to suggest releasing such technology to the general public.
“Thankfully I think it’s still many years away before such an aspiration can be achieved in any meaningful way, so we have time to put the regulation systems in place. But it is a matter of public safety that we progress this work with some urgency,” she told the Guardian.
The UK held an international summit on AI safety last year – attended by Meta – which committed to allowing governments to vet AI tools before and after their release.