New AI Safety Board gets power to block releases—but can keep findings secret.
S. 5576 — Artificial Intelligence Risk Management and Security Act of 2026 · Filed by Mark Warner (D-VA) · 2 cosponsors · Introduced Sep 29, 2026 · Referred to committee
Your members of Congress
Enter a ZIP to see where your representative and both senators stood on this bill.
Looked up on this device — your ZIP is never stored on our servers.
What it does
This bill creates a permanent Artificial Intelligence Safety Board within the Department of Commerce to develop technical standards and evaluate risks from advanced AI systems, particularly those capable of autonomous action. Developers of frontier AI models must comply with Board-adopted standards, submit safety plans, and provide pre-release access to the Board for testing. The bill also establishes a database for tracking AI security and safety incidents, with mandatory reporting requirements for developers and critical infrastructure operators.
Why we flagged it
The bill's core mechanism is establishing a government board to set and enforce technical standards for frontier AI systems before public release. This is regulatory oversight, not deregulation or subsidy—the operative verb is 'establish' and 'adopt,' and the direction of power flows toward government agencies and away from unilateral developer discretion.
What the text implies
- Closed-session authority (section 3(5)(F)) permits the Board to withhold safety assessments, incident reports, and vulnerability data from public view indefinitely under 'proprietary business information' exemption, potentially shielding developers from accountability even as citizens bear deployment risks.
- Pre-release access requirement (section 3(f)) gives the Board 45 days to evaluate frontier models before public release, but the bill contains no explicit authority to block release or impose conditions—the Board may only advise, creating a compliance theater risk if developers ignore findings.
- Mandatory reporting threshold (section 5(a)(3)) requires developers to report incidents within 30 days (or 72 hours for national security threats), but 'confirmation' is undefined—developers may delay reporting by disputing whether an incident has been 'confirmed,' creating a loophole.
- International participation clause (section 3(a)(2)(C)) permits the Secretary to negotiate agreements with foreign partners on AI safety standards and 'regulatory reciprocity,' potentially harmonizing U.S. standards downward to match weaker foreign regimes without congressional approval.
- Conflict-of-interest policy (section 3(a)(5)(D)) requires Board members to recuse from votes where they or their employer would 'directly and materially benefit,' but AI industry experts on the Board may benefit indirectly through industry-wide standard-setting that raises barriers to entry for competitors, creating a structural capture risk.
Section numbers refer to the bill text the analysis read — linked under Primary records below.
Who it affects
Citizens gain transparency and safety oversight of frontier AI systems before public release, with mandatory incident reporting and a public database of flaws. However, compliance costs may increase AI product prices, and the Board's closed-session authority and proprietary-information exemptions limit public visibility into actual safety assessments and incidents.
Who stands to gain
- Large AI developers (compliance cost barriers to entry for smaller competitors)
- Cybersecurity and testing service providers (Board may contract for secure environments and evaluati
- NIST and National Laboratories (staff support and secure computing environment contracts)
Named in the bill
Department of Commerce, Artificial Intelligence Safety Board, National Institute of Standards and Technology (NIST), Cybersecurity and Infrastructure Security Agency (CISA), National Security Agency (NSA), Department of Treasury, National Laboratories (Department of Energy), Director of National Intelligence, Attorney General, Frontier artificial intelligence model developers, Critical infrastructure operators
Where it stands
2 cosponsors: 2 Democrats.
- Sep 29, 2026 — Introduced · Congress.gov: “Introduced in Senate”
- Sep 29, 2026 — Referred to Senate Committee on Commerce, Science, and Transportation · Congress.gov: “Read twice and referred to the Committee on Commerce, Science, and Transportation”
Dates and quoted wording are Congress.gov's action record; the timeline shows status changes, not every procedural step.
How this was measured
Analysis — Quorum's AI read the bill text published by Congress.gov (32,983 characters) on Oct 13, 2026. Section numbers in the findings refer to that text, linked below; transparency and hidden-provision scores are compared against the median of 16,222 analysed bills.
Status and sponsors — Congress.gov's bill record — actions, committee referrals and cosponsors — loaded nightly. The timeline shows status changes, not every procedural action.
As of — page rendered 2026-10-20.
“New AI Safety Board gets power to block releases—but can keep findings secret.” QuorumCivic. https://share.quorumcivic.app/bill/119/s5576 Report an error
The week in bills
One email, Sunday: what Congress passed, what it changed, and who paid.
One link arrives to confirm the address. It is used for this email and nothing else, and every issue carries an unsubscribe link.