The Senate Commerce Committee voted 14-12 today to advance the Artificial Intelligence Accountability and Transparency Act (AIATA), marking the first major federal legislation to address AI training data rights and commercial AI system oversight.

The bipartisan bill, co-sponsored by Senators Maria Rodriguez (D-CA) and James Chen (R-TX), would require AI companies to disclose the sources of their training data and establish a framework for content creators to receive compensation when their work is used in commercial AI systems.

Key Provisions

The legislation includes several groundbreaking requirements for AI companies operating in the United States:

• Training Data Transparency: Companies must publish detailed reports about the sources and licensing status of data used to train AI models
• Creator Compensation Framework: Establishment of a federal registry for content creators to claim compensation for AI training usage
• Audit Requirements: Annual third-party audits of AI training practices for companies with models over 100 billion parameters
• Real-time Attribution: AI systems must provide source attribution when directly quoting or closely paraphrasing copyrighted content

Industry Response

The technology industry remains divided on the legislation. Major AI companies have expressed concerns about the compliance burden, while publisher and creator organizations have largely supported the framework.

“This legislation strikes the right balance between innovation and creator rights,” said Dr. Sarah Kim, Executive Director of the Digital Publishers Alliance. “AI companies should not be able to build billion-dollar businesses on the backs of creators without fair compensation.”

However, TechForward, a trade association representing major AI companies, argued that the requirements could stifle innovation. “These reporting requirements are overly burdensome and could push AI development overseas,” said spokesperson Michael Torres.

Next Steps

The bill now moves to the full Senate for consideration. Senate Majority Leader Patricia Williams indicated that floor time could be scheduled within the next month, depending on the legislative calendar.

The House has been working on parallel legislation, though their version includes more stringent requirements around algorithmic auditing and bias testing. Congressional staffers suggest that any final bill would likely need to reconcile differences between the two chambers.

If enacted, the legislation would take effect 18 months after signing, giving companies time to build compliance systems. The Federal Trade Commission would be tasked with enforcement, with penalties ranging from warnings to fines up to $100 million for repeated violations.