California's AI Transparency Law Failed: State Agencies Used Six High-Risk Systems Undetected
Key Takeaways
- ▸California's transparency law (AB 302) failed to identify six high-risk automated decision systems in its first report due to lack of verification mechanisms
- ▸The law depends entirely on voluntary agency self-reporting with no penalties for non-compliance or false reporting
- ▸State AI systems make life-consequential decisions about public benefits, housing, and healthcare without public awareness
Summary
California's Assembly Bill 302, signed three years ago, required the state's Department of Technology to conduct a comprehensive inventory of all high-risk automated decision systems used by state agencies and publish yearly findings. In its 2025 report, the department claimed California state government used no high-risk AI systems—a conclusion researchers quickly disproved by filing a public records request that revealed the initial response was a single spreadsheet with "no" listed for every agency.
Following this embarrassment, the state agency interviewed a few departments that finally disclosed they had been using at least six automated decision systems to make consequential government determinations about Californians' lives. These systems mine sensitive personal data to automate decisions about cash assistance eligibility, housing qualifications, and medical care access.
The law's fundamental failure stems from its complete reliance on self-reporting with no verification process or enforcement penalties for agencies that withhold information. Researchers from UCLA's Center on Resilience and Digital Justice and Georgetown Law's Center on Privacy and Technology argue this exposes a broader problem: transparency-only regulation without verification mechanisms cannot fulfill even basic public accountability, let alone drive meaningful change.
- Transparency-only regulation may be insufficient to regulate government AI without enforcement and verification requirements
