A bipartisan group of senators, including U.S. Senator Katie Britt, R-Alabama, introduced legislation last week that would require new transparency measures for artificial intelligence companies.
The AI Systems Transparency Act, or ASTA, is sponsored by Britt and U.S. Senators Chris Coons, D-Delaware; James Lankford, R-Oklahoma; and Brian Schatz, D-Hawaii.
The legislation would require AI companies to publicly disclose the data their models collect from users, safeguards they have in place for children and adults, and guardrails they have enacted to stop systems from “going rogue” or being exploited by bad actors.
Under ASTA, the Federal Trade Commission would have the power to enforce the transparency requirements.
“Americans are understandably concerned with the rapid development of AI, and they deserve straight answers,” Britt said in a written statement released by her office Monday.
“We know that this technology is not slowing down, which is why it’s critical that AI companies provide transparency to both researchers and parents on how their systems work and address potential risks – especially risks related to America’s youth,” Britt said. “I’m proud to join my colleagues in this effort to provide the American people visibility into how these systems are built, tested, and secured.”
ASTA would apply to closed- and open-source AI models.
The bill would establish criteria for the size and scope of companies over which the FTC would have jurisdiction. Covered models would include open-source AI models that have or are expected to have more than 1 million users per month for most of the year, as well as models that allow minors to create accounts.
The bill would require companies to disclose information about the structure and policies typically included on “model cards,” brief documents that explain an AI model’s purpose, training data, performance metrics and known limitations.
Companies also would have to disclose preventive safeguards they use for child safety, mental health, privacy, cybersecurity, disaster risk and autonomous loss of control.
“AI is rapidly evolving in ways that are radically reshaping our society and transforming our world and how we interact with each other, and Americans are rightfully concerned about what AI companies and their government are doing to protect against worst case scenarios,” Coons said. “Companies must not be allowed to hand wave away these doubts without meaningful disclosures, safeguards, and oversight in place. Americans deserve to know what they are actually doing to place safety guardrails on large AI models, and that’s what this bipartisan bill will ensure.”
“Trust has to be earned, and right now families are being asked to trust AI companies with very little information,” Lankford said. “The AI Systems Transparency Act gives parents and consumers the straightforward answers on how these AI models are built and how they interact with America’s children. Transparency on innovation earns the public’s trust.”
The legislation also would require companies to disclose information related to common policy violations and update all disclosures with each new model or substantial update to an existing model. Required disclosures would include both a consumer-facing disclosure and one for use by independent researchers and evaluators.
“Recent reports of security incidents with AI models make transparency and oversight all the more urgent. Our bill requires disclosures that help both consumers and evaluators understand both the capabilities of and risks posed by these models and what companies are doing to keep people safe,” Schatz said.
ASTA has received endorsements from Americans for Responsible Innovation, a nonprofit focused on AI governance, and Encode AI, a nonprofit that advocates for AI technology regulation and governance.
“Americans are increasingly relying on AI in their everyday lives, but the public has no visibility into how those systems are built and tested. We cannot rely on the word of AI companies alone that their products are safe – there must be evidence to back that up,” said Brad Carson, president of Americans for Responsible Innovation. “This bill will shed light on the proactive safety measures and potential risks of AI models so users can decide for themselves whether these systems are trustworthy.”
“People deserve to know what AI companies are doing to keep them safe,” said Sunny Gandhi, co-executive director of Encode AI. “The AI Systems Transparency Act would require meaningful public disclosures about AI capabilities and safeguards, including whether developers have evaluated serious risks such as loss of control.”
The legislation’s introduction follows several letters led by the bill’s sponsors in December 2025 urging OpenAI, Microsoft, Google, Anthropic, Meta, Luka, Character.AI and xAI to improve transparency about their models’ capabilities and the risks they pose to users.
