AI Flaw Reporting and Security Enhancement Act

Jun 18, 2026

Summary

Sets up a system for people and companies to voluntarily report problems with AI to help make it safer and more secure for everyone.

What problem does this solve?

There is no standard way to report or track when AI systems make mistakes or have security problems. This bill creates a national program and database so experts can share information about AI flaws and fix them faster.

What does this bill do?

Creates a voluntary AI flaw reporting program

Directs the National Institute of Standards and Technology (NIST) to create a program for the voluntary reporting, collection, and tracking of flaws in artificial intelligence systems.

Establishes common definitions for AI flaws

Requires NIST to work with experts to create common definitions for AI issues like vulnerabilities, failures, accidents, and misuse to ensure everyone uses the same language.

Develops a national database of AI flaws

Tasks NIST with developing or funding the creation of a national database to serve as a central place for tracking reported AI flaws.

Supports new standards and best practices

Supports the creation of technical standards, guidelines, and best practices for managing, reporting, and fixing AI flaws across different industries.

Requires a report to Congress

Mandates that the Director of NIST submit a report to Congress within three years on the program's progress, including findings and recommendations.

Who does this affect?

AI developers and technology companies
Academic and non-profit researchers
Federal agencies using AI

What is the real world impact?

•

Improves AI safety and security

Creates a central system for finding and fixing problems in AI. This helps prevent accidents, misuse, or security breaches before they happen by letting experts share information.

•

Establishes industry standards

Encourages companies, researchers, and the government to agree on common definitions and best practices for AI. This helps everyone work together to build more reliable AI systems.

•

Relies on voluntary participation

The program is voluntary, which might not be strong enough to ensure all important flaws are reported. Some companies may choose not to share information about their AI systems to avoid bad publicity or legal issues.

When does this start?

This bill would take effect when it becomes law, with a specific deadline for a report to Congress.

Report to Congress

The Director of NIST must submit a report on the program's implementation no later than three years after the bill becomes law.