What happened
On 24 September 2026 Senators Chris Coons (D-DE), Katie Britt (R-AL), Brian Schatz (D-HI) and James Lankford (R-OK) introduced bipartisan legislation (reported as the AI Systems Transparency Act) that would require AI companies to disclose more information about their models, user data collection and safety measures — including safeguards addressing child safety, mental health, privacy, cybersecurity and risk of systems operating beyond human control. The FTC would enforce the requirements. It applies to both closed- and open-source models, with criteria determining which companies fall under FTC oversight, and requires updating disclosures when releasing a new model or making substantial changes.
Why it matters
It is the window's flagship bipartisan Senate AI transparency bill and would, if enacted, create FTC-enforced disclosure duties for AI model providers on safety and data practices, with downstream effects on how deployers assess model risk. Watch-list tier pending movement.
Action needed
AI companies operating in the US should monitor the bill; if enacted, FTC-enforced transparency disclosures on model structure, safety safeguards, and data practices would apply to qualifying AI systems on a sliding scale by size.