HR-9333 — AI Flaw Reporting and Security Enhancement Act

YourVoice.Now Summary

Transparency & Accountability

A new federal database would track flaws in AI tools — but no company would have to report one.

Transparency & Accountability

National database of AI flaws — reporting would be voluntary

NIST would set up one place to collect reports of AI flaws. Nobody would have to file one, and a college or research institute could run the database instead.

Public information about AI flaws — shared ways to report them

The program would build shared ways to report AI flaws, some of them automatic. It would also set standards for when a flaw goes public, and when it waits.

More about this bill

Nothing in your daily life would change right away. The government would start to collect reports about flaws in AI tools. That means security holes, safety failures, accidents, and misuse. Reporting would be up to each company. Nothing in the bill forces a company to report a flaw. NIST, the agency that writes tech standards for the government, would run the program. It would work with CISA, the agency that guards against computer attacks. The two would bring in companies, colleges, nonprofits, and groups that speak for the public. Together they would agree on what counts as an AI flaw. They would also rank how serious each flaw is. Another goal is finding flaws faster. NIST would set up a national database of these flaws, or add them to one that already exists. A college or research institute could run that database instead. NIST would report to Congress on its work within three years.

Congressional Summary

AI Flaw Reporting and Security Enhancement Act This bill establishes a program to facilitate the voluntary reporting and tracking of artificial intelligence (AI) flaws, to be administered by the National Institute of Standards and Technology (NIST). In carrying out this program, NIST must seek to convene various stakeholders to establish common definitions for terms related to AI flaws and criteria for the classification of AI flaws (e.g., security-related flaws and safety-related flaws). The group must also support the development of technical standards and guidance related to detecting, managing, and disclosing AI flaws and prioritizing the remediation of such flaws. Further, NIST must develop, or enter into cooperative agreements with institutions of higher education or research institutions to develop, infrastructure for the voluntary reporting, collection, and tracking of AI flaws. This must include a national database of AI flaws or the modification of an existing national database to account for AI flaws. (NIST currently administers a national database of cybersecurity vulnerabilities.) NIST must consider certain topics when developing this infrastructure, including the interoperability of the infrastructure with relevant existing systems, standards, and best practices. Within three years of the bill’s enactment, NIST must report to Congress on the implementation of these provisions. Under the bill, an AI flaw is a set of conditions or behaviors that allow for the violation of certain policies (e.g., safety or security policies) and is not necessarily associated with malicious intent.

Legislative Subjects

Computer security and identity theftComputers and information technologyConsumer affairsData collection, sharing, protectionPerformance measurementProduct safety and qualityTechnology assessment

Details

Congress: 119th
Chamber: House
Status: summarized
Action: Introduced in House
Action Date: 2026-06-18
Date Added: 2026-07-25
Source: Congress.gov →

Like reading a bill in plain English?

We're building an app that does this for every bill in Congress and lets you tell your reps how you want them to vote. We're a small team getting ready to launch, and we're trying to show investors that real people want this. Be one of them. Help us get it built. Leave your email and we'll tell you the moment the app is ready.