The United States Senate passed the Artificial Intelligence Accountability Act by a 71-29 bipartisan vote, establishing for the first time a clear federal legal framework for AI liability. The bill now heads to the President's desk after the House passed it last month.
Under the new law, companies that deploy AI systems in high-risk contexts β healthcare, criminal justice, hiring, lending, and critical infrastructure β must conduct mandatory pre-deployment safety testing, maintain audit logs, and carry liability insurance. If an AI system causes harm, the deploying company bears legal responsibility unless it can demonstrate it followed approved safety standards.
The bill also creates an AI Safety Board within the FTC with authority to investigate AI harms, issue fines of up to $50 million per violation, and mandate system recalls for AI products deemed unsafe.
Tech industry groups called the law "workable" after months of intense lobbying softened several provisions. Civil rights organizations praised the bill's explicit ban on AI systems that exhibit racially discriminatory outcomes in lending and hiring.
