Texas House bill would require human override and termination controls for high-risk AI systems that act on their own

Texas Legislature · Bill text

Read the source (opens in a new tab)
What happened
Texas Representative Bhojani filed HB 510 for the Legislature's 90th regular session on November 9, 2026. It is a proposal and has not passed.
Key points
  • It covers AI systems "capable of independently initiating or executing an action" that risks substantial harm to people's health or safety, property, critical infrastructure, or information security.
  • Deployers would have to let a trained human operator monitor the system and "approve, reject, modify, override, and terminate" its decisions or actions, and keep mechanisms reasonably designed to prevent the system from circumventing those controls.
  • Developers would supply and test the controls before deployment and after material changes; deployers would test them at least yearly and after any material failure before resuming operations.
  • The attorney general would enforce it with civil penalties and could seek an emergency injunction when a system presents "an immediate and substantial risk of serious harm"; whistleblowers are protected from retaliation.
Evidence
Bill text as introduced, with the Legislature's bill history showing it filed on November 9, 2026. It shows what the bill would do, not whether it will advance.
Uncertain
The bill needs committee action, passage by both chambers and the governor's signature. If enacted as filed, it would take effect September 1, 2027.

See it in the latest record