Definition
The emerging set of laws and rules that would require AI-model developers to disclose how their models work, what they were trained on, and what risks they pose — typically to a regulator or the public, on a standardised schedule. Bipartisan US Senate and House bills and California's implementation roadmap are converging on mandatory reporting for frontier and foundation models.
Why it matters
Mandatory disclosure is how regulators compensate for being unable to inspect models themselves; enterprises that buy or build models will inherit these reporting and audit obligations down the chain.