New legislation introduced by US lawmakers aims to simplify the process of limiting exports of AI models | Technology News

A bipartisan group of lawmakers introduced a bill aimed at protecting U.S. AI technology from foreign threats, particularly from China and Russia. The bill, sponsored by both Republicans and Democrats, would make it easier for the Biden administration to impose export controls on AI models that pose risks to national security.

The legislation, crafted with input from administration officials, seeks to prevent U.S. adversaries from using advanced AI models to launch cyber attacks or develop biological weapons. It also gives the Commerce Department the authority to regulate Americans working with foreigners on potentially dangerous AI systems.

The move comes as concerns grow over China’s reliance on open source AI models developed in the West, such as Meta Platforms’ “Llama” series. Chinese AI companies have faced backlash for using these models, with some even being accused of building their own AI systems on top of them.

In a separate development, Microsoft’s $1.5 billion investment in a UAE-based AI firm has raised eyebrows in the U.S. amid worries about deepening ties between China and Gulf states. The deal, which includes a security agreement with both the U.S. and UAE governments, allows the AI firm to use Microsoft cloud services for its applications.

The bill aims to address these concerns by giving the Commerce Department more power to regulate the export of AI models, particularly open source ones that can be freely downloaded. If approved, the measure could help safeguard U.S. AI technology and prevent it from falling into the wrong hands.