U.S. House Foreign Affairs Committee Approves Export Restrictions on AI Systems

The U.S. House Foreign Affairs Committee voted overwhelmingly to advance a bill that simplifies export restrictions on artificial intelligence systems for the Biden administration. The decision stems from concerns about China's potential misuse of these systems to enhance its military capabilities. The bill, introduced by Republican representatives Michael McCaul and John Molinar, along with Democrats Raja Krishnamoorthi and Susan Wild, grants the U.S. Department of Commerce the authority to prohibit Americans from collaborating with foreigners on developing AI systems that pose national security risks. McCaul, who chairs the committee, warned that without this law, 'large American AI companies could inadvertently help advance China's technology to bolster its military ambitions.' He added, 'As China seeks to expand technological advancements to enhance its surveillance system and war machine, it is crucial to protect our sensitive technologies from falling into their hands.' Currently, there are no barriers for major AI producers in the U.S., such as Microsoft's OpenAI, Alphabet's Google DeepMind, or Anthropic, allowing them to sell their powerful AI models to anyone in the world without government oversight. The new bill passed with 43 votes in favor and only 3 against, but it must first be approved by the House of Representatives and the Senate before being signed into law by the President. This bill indicates that Washington is prepared to curb China's ambitions in AI, driven by fears that Beijing could use it for interfering in other countries' elections, creating biological weapons, or conducting cyberattacks. A bipartisan group of U.S. senators, including Senate Majority Leader Chuck Schumer, also urged Congress last week to approve $33 billion in funding for AI research to keep the U.S. ahead of China in technology. Reuters previously reported that any potential export controls would likely target Iran, Russia, China, and North Korea. Microsoft reported last February that hackers backed by Iran, Russia, China, and North Korea were using tools from 'OpenAI,' supported by the company, to enhance their skills and deceive regarding their objectives. Microsoft stated that Iranian government-affiliated hackers relied on large language models (AI systems for understanding, generating, and responding to human language) to write more convincing emails and at one point used them to draft a message attempting to lure 'known feminists' into a trap by directing them to a website. According to Reuters, U.S. officials expressed their concerns about China's 'misuse' of AI during their first official bilateral talks on the matter in Geneva. The White House: Recent hacking operations by the Iranian government in the U.S. are a fresh warning to strengthen cybersecurity. Axios: Iranian-backed hackers have increased their activities in the region. An American cybersecurity company states: IRGC hackers are seeking money. The U.S. urged China and Russia not to entrust 'AI' with nuclear weapons control. Blinken urged China to prevent Iran from escalating tensions in the region. The Japanese Prime Minister told American lawmakers: U.S. leadership in global affairs is 'essential.' The U.S. blacklisted several Chinese and Russian entities due to their connection with Iran's drone program. Microsoft: China uses fake accounts to alter important election results worldwide in its favor. The FBI director warned about the use of AI for interference in U.S. elections.