AI Accountability Act

This proposal, called the Artificial Intelligence Accountability Act, would have the National Telecommunications and Information Administration (NTIA) study how to make AI systems more trustworthy and safe, especially in areas like telecom networks and social media. It would also require public meetings to gather input and recommend what information about AI systems should be shared with people, communities, and businesses, and the best ways to share it. The study must look at how these checks can help close the digital divide, improve digital inclusion, and cut risks like cybersecurity threats. “Accountability measures” include things like audits, assessments, and certifications that show an AI system can be trusted. Within 18 months of the law taking effect, NTIA would report findings and recommendations to Congress, including what information should be available and how to make it available.

Key points

• Who is affected: People using or impacted by AI systems; communities; businesses; and stakeholders like industry, academics, and consumers who can give input at public meetings.
• What changes: NTIA must study AI accountability in communications technologies, hold public meetings, and recommend what AI information should be shared and how; it defines “accountability measures” as audits, assessments, or certifications .
• When: Reports to Congress are due within 18 months after the law takes effect.