Search on TFTC
US Bill Proposes Harsh Limits on AI Exports

US Bill Proposes Harsh Limits on AI Exports

May 10, 2024

US Bill Proposes Harsh Limits on AI Exports

A new legislative proposal aimed at restricting the export of artificial intelligence (AI) technology has been introduced by a group of bipartisan lawmakers. The bill, sponsored by Representatives Michael McCaul (R), John Molenaar (R), Raja Krishnamoorthi (D), and Susan Wild (D), seeks to make it easier for the Biden administration to enforce export controls on AI models.

The proposed legislation would grant the Commerce Department explicit authority to prevent Americans from collaborating with foreign nationals in the development of AI systems deemed dangerous to U.S. interests. This move is designed to reinforce any potential AI export regulations against legal challenges and was informed by input from officials within the Biden administration.

Despite this initiative, the Commerce Department and White House have not yet made any public comments regarding the bill. Reuters has reported that the United States is planning to establish new export controls on cutting-edge proprietary AI models as a protective measure against China and Russia.

Under the current U.S. law, the Commerce Department faces challenges in regulating the export of open-source AI models, which are readily available for download. The new bill, if passed, would amend the International Emergency Economic Powers Act, eliminating barriers to the regulation of open-source AI exports, as well as providing the Commerce Department with clear regulatory authority over AI systems.

China's reliance on open-source AI models, notably those developed by companies like Meta Platforms, has been substantial. The Beijing Academy of Artificial Intelligence acknowledged in March that most Chinese AI models were built on Meta's Llama models, highlighting a significant obstacle to China's independent AI development.

Furthermore, the bill's introduction follows recent events where 01.AI, a prominent Chinese AI company led by former Google executive Lee Kai-fu, faced criticism after it was discovered that its Yi-34B AI model was based on Meta’s Llama system. This revelation raised questions about the origins and integrity of Chinese AI advancements.

In related news, Microsoft announced a $1.5 billion investment in the United Arab Emirates-based AI firm G42, which includes an agreement for G42 to use Microsoft's cloud services for its AI applications. This deal, which involved security agreements with both the U.S. and UAE governments, came amidst heightened U.S. concerns about the growing connections between China and Gulf states, including the UAE.

As the bill progresses through the legislative process, its future implications for U.S. technology exports and international collaborations in AI development remain to be seen. The proposed measures may significantly impact how AI technology is shared globally.


Current Block Height

Current Mempool Size

Current Difficulty