A bipartisan bill that would support the Biden administration in enforcing export restrictions on the nation’s best artificial intelligence (AI) models has been unveiled by American legislators. The plan, which would prevent future regulations on AI exports from facing legal challenges, was unveiled late on May 8 by House Republicans Michael McCaul and John Molenaar and Democrats Raja Krishnamoorthi and Susan Wild. The Commerce Department would also be granted “express authority” to prevent individuals of the United States from working with foreign agents in the creation of AI systems that could endanger national security.
The U.S. Commerce Department, which is in charge of export policy, finds it more challenging to regulate open-source AI models under the current laws. The newly proposed bill will reduce obstacles to controlling open source AI exports if it is approved. This measure comes after a Reuters story published on Wednesday stating that the United States is prepared to defend its in-house AI against threats from China and Russia. It declared that it will put export restrictions on its most sophisticated in-house AI models.
China has been known to utilize open-source models originating from the U.S., like Meta’s Llama models, to develop its own AI.
It also comes after Microsoft and G42, a Dubai-based artificial intelligence startup in the United Arab Emirates (UAE), recently reached an agreement on April 16.Microsoft promised to support local AI innovation in the UAE with a $1.5 billion partnership. The UAE will now have access to state-of-the-art AI technologies created by Microsoft in the United States thanks to this agreement. This is only one of several agreements Microsoft has with other countries about the creation and application of its artificial intelligence models and cloud computing services.
The United States’ primary concern, though, remains the possibility of China gaining access to advanced AI technology.The AI model Yi-34B of 01.AI, a well-known Chinese startup run by Lee Kai-fu, a former Google executive, was developed with the aid of Meta’s Llama system, the company disclosed in November 2023. The US government has taken a number of actions to punish China and limit its access to AI technology produced in the US. The People’s Liberation Army (PLA) was accused of using Chinese AI startup Baidu’s technology for military purposes, which put the company in danger of penalties in January.