US lawmakers unveil bill to make it easier to restrict exports of AI models

A bipartisan group of lawmakers unveiled a bill late Wednesday that would make it easier for the Biden administration to impose export controls on AI models, in a bid to safeguard the prized US technology against foreign bad actors.

The bill, sponsored by House Republicans Michael McCaul and John Molenaar and Democrats Raja Krishnamoorthi and Susan Wild, would also give the Commerce Department express authority to bar Americans from working with foreigners to develop AI systems that pose risks to US national security.

The legislation aims to bulletproof any future AI export regulations from legal challenges and was crafted with input from Biden administration officials. It comes as worries mount that US adversaries could use the models, which mine vast amounts of text and images to summarize information and generate content, to wage aggressive cyber attacks or even create potent biological weapons.

The Commerce Department and White House did not immediately respond to requests for comment on the legislation.

Reuters reported on Wednesday that the United States is poised to open up a new front in its effort to safeguard US AI from China and Russia with preliminary plans to impose export controls on the most advanced proprietary AI Models.

But under existing US law, it is much more difficult for the Commerce Department, which oversees US export policy, to regulate the export of open source AI models, which can be freely downloaded.

If approved, the measure would remove roadblocks to regulating the export of open source AI contained in the International Emergency Economic Powers Act and also give the Commerce Department express authority to regulate AI systems.

China has been heavily relying on many open source models developed in the West such as Meta Platforms' "Llama" series.

In March, the Beijing Academy of Artificial Intelligence, a high-level research lab, was quoted by Chinese state media as stating that the majority of homegrown Chinese AI models were in fact built using Meta's Llama models and that this posed a key challenge to China's AI development.

In November 2023, 01.AI, one of the most high-profile AI unicorns in China founded by Google's former executive Lee Kai-fu, faced a major backlash after some AI engineers found that its AI model Yi-34B was built on Meta's Llama system.

It also comes after Microsoft announced it was investing $1.5 billion in United Arab Emirates-based artificial intelligence firm G42, giving G42 permission use to use Microsoft cloud services to run its AI applications.

The deal, which involved a security agreement inked with the US and UAE governments, was unveiled despite growing concerns in the United States over deepening ties between China and the Gulf states, including the UAE.