POLISCOPE
Back to feed
SB 833CALIFORNIASession 20252026
High Impact

California requires state agencies to implement human oversight for AI systems in critical infrastructure

Original title: Critical infrastructure: artificial intelligence systems: human oversight.

August 13, 2026

Track this bill to get notified when it advances a stage. One tap to stop, anytime.

The Frame

What this does

This bill creates a mandatory safety layer for state-managed , ensuring that automated AI decisions cannot be implemented without human verification, thereby shifting liability and operational control to designated oversight personnel.

Who is mentioned in the record

Potentially affected actors named in the source documents. Mention is not a position.

California state agencies

Agencies must implement new oversight mechanisms, training, and annual reporting requirements for AI systems.

Department of Technology

The department is tasked with creating new AI safety training programs and collecting assessment data from other agencies.

What changed

Last recorded activity August 13, 2026.

What's next

Introduced.

Summary

State agencies operating must establish real-time human oversight for any artificial intelligence systems they deploy by July 1, 2026. This mandate requires human personnel to review and approve all AI-proposed actions before they are executed and mandates annual safety assessments.

Key Facts

You don't have to trust us. Each fact below is taken straight from the official document - click any one to see the exact passage, highlighted in the original.

Why It Matters

This bill creates a mandatory safety layer for state-managed , ensuring that automated AI decisions cannot be implemented without human verification, thereby shifting liability and operational control to designated oversight personnel.

Frequently Asked Questions

Does this bill apply to private companies?
No, the bill specifically defines an 'operator' as a state agency responsible for .
What happens if an AI system proposes an action?
Under this bill, a human must review and approve the proposed action before it can be executed.

News Coverage

No news coverage found yet. Articles are indexed twice daily.

Discoveries

Patterns POLISCOPE noticed across the record. These are observations to investigate, not conclusions.

policy shift90% confidence

Mandatory Human-in-the-Loop

The bill represents a shift from voluntary AI safety guidelines to a mandatory real-time approval requirement for state infrastructure.

Connected Entities

organizationOffice of Emergency ServicesOversees the California Cybersecurity Integration Center.Map →
organizationCalifornia Cybersecurity Integration CenterExisting hub for state cybersecurity activities.Map →
organizationDepartment of TechnologyResponsible for developing AI safety training and receiving assessment summariesMap →

Sources

Open source document

openstates.org

Analysis Score

0–100
  • Significance85
    How much this matters to a regular citizen
  • Controversy40
    Intensity of disagreement among stakeholders
  • Entertainment10
    Compellingness for a non-policy-wonk reader
  • Buzz50
    Current news / social attention level

Publisher tools

Share or embed this record

POLISCOPE publisher tools

Share or embed this record