HR 10180
Self-Improving AI Monitoring Act
TL;DR
Rep. George Whitesides (D-CA) introduced this bill to have the federal government monitor and study AI systems that can improve themselves without human intervention (think AI models that rewrite their own code or retrain themselves). It's currently sitting in the House Science Committee and focuses on tracking risks from recursive self-improvement, not regulating specific business practices yet.
How This Might Impact Your Business
AI labs and foundation model developers (OpenAI, Anthropic, Google DeepMind, Meta AI) would face new federal monitoring of self-improving or recursively self-training systems.
No immediate compliance requirements or penalties for typical enterprise AI users, this is a monitoring and reporting framework, not a restriction.
Companies deploying autonomous AI agents that modify their own code or retrain themselves in production may eventually need to report capabilities to a federal body.
Cloud providers hosting large training runs (AWS, Azure, Google Cloud) could be pulled into future reporting obligations tied to compute thresholds.
Defense contractors and national security vendors using advanced AI should expect closer federal scrutiny of autonomous system capabilities.
Bill is at the earliest stage (committee referral), meaning language could change significantly before any vote, and passage this session is uncertain.
No specified fines, deadlines, or small business exemptions in the current version since the bill's focus is establishing federal visibility, not enforcement.
What Should You Do
Ask your CTO or head of AI whether any deployed systems qualify as 'self-improving' (automated retraining, agentic code modification, reinforcement learning in production) and document those capabilities now.
Assign someone in government affairs or legal to track this bill through the House Science, Space, and Technology Committee for markup activity.
If you're a frontier AI developer or major cloud provider, engage with committee staff early to shape definitions of 'self-improving AI' before they harden.
Brief your board on the trend: federal interest in autonomous AI monitoring is accelerating, and reporting obligations are the likely first step before harder regulation.
Review internal AI governance documentation to ensure you can quickly produce capability reports if future rules require disclosure.
Who It Affects
Sponsors
Status Timeline
committee
Referred to the House Committee on Science, Space, and Technology.
August 27, 2026
committee
Referred to the House Committee on Science, Space, and Technology.
August 27, 2026