Sets Standards Against AI Cyber Attacks
HR10873-119 · Introduced Oct 22, 2026
Why it matters
Companies that build or use AI could get voluntary guidance to defend against AI-based attacks. Participation and following the standards would be optional.
What it does
- Requires NIST to seek work with private AI companies to define AI distillation attacks within 180 days
- Asks NIST to publish voluntary standards for detecting and preventing those attacks within one year
- Creates an AI Cybersecurity Roundtable within 180 days to develop voluntary defensive AI standards
- Asks the roundtable to include AI companies and owners of critical infrastructure
- Excludes foreign companies tied to China, Russia, Iran, North Korea, or Cuba from the definition of stakeholders
In committee
Next: a committee vote
- Introduceddone
- Committeecurrent step
- Housenot yet
- Senatenot yet
- Presidentnot yet
About this bill
Would direct NIST to create voluntary standards to detect and prevent AI distillation attacks and defensive AI cybersecurity uses. Work would begin within 180 days, with standards published within one year.
AI Cyber Defense Act
Perspectives
💬 The case for
Cybersecurity groups would likely say clear standards help companies defend against AI-driven attacks. Voluntary guidance avoids forcing new rules on businesses.
💬 The case against
Some AI companies might say voluntary standards do little without enforcement. Others may worry about government shaping the industry's security practices.
Sponsored by
Sponsor details for this federal bill are at the source link.
News
Loading coverage…