Meta's Zuckerberg puts faith in open-source AI to maintain US lead over China
Meta Platforms chief executive Mark Zuckerberg, whose company runs Llama, the world's largest open-source large language model (LLM), said artificial intelligence based on open-source technology will help the US maintain a lead over China.
"Some people argue that we must close our models to prevent China from gaining access to them, but my view is that this will not work and will only disadvantage the US and its allies," Zuckerberg wrote in a blog post published on July 23, advocating for open source AI and announcing its own open source model Llama 3.1.
Having only closed models will result in a world where leading models can only be accessed by a handful of large companies and the US' geopolitical adversaries, who are "great at espionage", the Facebook founder argued. "Stealing models that fit on a thumb drive is relatively easy, and most tech companies are far from operating in a way that would make this more difficult."
Do you have questions about the biggest topics and trends from around the world? Get the answers with SCMP Knowledge, our new platform of curated content with explainers, FAQs, analyses and infographics brought to you by our award-winning team.
The Meta chief's comments come amid a fierce AI race between China and the US. Driven by national security concerns, Washington has intensified measures preventing China's access to advanced chips and curbing American investment in Chinese AI projects. In response, Beijing has urged domestic firms to build the country's own models independent of foreign technology.
Many Chinese developers still rely on Meta's Llama models. Photo: Shutterstock Images alt=Many Chinese developers still rely on Meta's Llama models. Photo: Shutterstock Images>
Many Chinese tech firms still rely on US AI models, including closed versions such as OpenAI's suite of GPT models, and Meta's open source Llama.
OpenAI, which sparked the generative AI craze with the release of ChatGPT in November 2022, last month said it would take additional measures to block access to its AI models via application programming interfaces (APIs) from unsupported regions, including mainland China and Hong Kong.
Some Chinese developers have scrambled to keep access to OpenAI models, by using virtual private networks and third-party services.
Many Chinese developers have also been relying on Meta's Llama models, an indication of the hurdles local companies face in trying to reach self-sufficiency, according to the Beijing Academy of Artificial Intelligence.
Still, Zuckerberg argues that open sourcing AI models will help the US maintain its advantage.
"There's one strain of thought, which is like, we need to lock it all down, and I just happen to think that's really wrong because the US thrives on open and decentralised innovation," Zuckerberg said in a video interview with Bloomberg published on Thursday. "That's the way our economy works. That's how we build awesome stuff."
But to preserve its lead, Washington should still work with leading American companies to make sure that the US has a "perpetual first mover advantage" on leading technologies for the purpose of national defence, Zuckerberg said in the interview.
The US should aim to maintain a six-to-eight-month lead over China, he added, arguing that it is an unreasonable goal to try and be five to 10 years ahead of its main rival.
This article originally appeared in the South China Morning Post (SCMP), the most authoritative voice reporting on China and Asia for more than a century. For more SCMP stories, please explore the SCMP app or visit the SCMP's Facebook and Twitter pages. Copyright © 2024 South China Morning Post Publishers Ltd. All rights reserved.
Copyright (c) 2024. South China Morning Post Publishers Ltd. All rights reserved.