# What the Artificial Intelligence Accountability Act Actually Does

This bill asks the federal government to study how artificial intelligence systems can be made more trustworthy and transparent, particularly when they're used by companies that provide internet and phone services or social media platforms. Right now, there's no clear standard for how AI should be checked or verified to make sure it's working fairly and safely. The bill directs a specific official at the Commerce Department to figure out what "accountability" should look like for AI—things like audits, testing, or certifications that prove an AI system is reliable and won't cause harm.

The study will look at several practical questions: How are companies currently building trust into their AI systems? Can better accountability measures help close the digital divide and get more people online? What security risks come with AI, and how can accountability help reduce them? The bill also wants to know what methods companies are already using and what obstacles they face in creating trustworthy AI. To get real-world input, the official must hold public meetings where tech companies, universities, and everyday consumers can share their thoughts and concerns.

Beyond studying accountability, the bill also requires a separate investigation into what information people should know when they interact with AI. The government will meet with industry representatives, academics, and consumers to figure out what details matter most to regular people and the best ways to share that information. For example, should you know when you're talking to an AI chatbot? Should companies disclose how an AI made a decision about you? After 18 months, the Commerce Department must report back to Congress with recommendations for both government and private sector actions to make AI systems more trustworthy and transparent.

Artificial Intelligence Accountability Act

Harder, Josh

Artificial Intelligence Accountability Act or the AI Accountability Act This bill requires the National Telecommunications and Information Administration (NTIA) to study and report on accountability measures for artificial intelligence (AI) systems. (Generally, AI refers to computerized systems that work and react in ways commonly thought to require human intelligence, such as certain systems that solve complex problems in real-world situations.) Specifically, the NTIA must study, solicit stakeholder feedback about, and report to Congress concerning mechanisms (e.g., audits, certifications, and assessments) to provide assurances that an AI system is trustworthy. Among other topics, the study, feedback, and report must address how the various accountability measures are currently incorporated into AI systems, their effectiveness, and barriers to their creation. In addition, the NTIA must consult with stakeholders through public meetings to develop recommendations about the information that should be available to consumers who interact with AI systems and the methods for making that information available.