California Bill to Regulate High-Risk AI Fails to Advance
A California artificial intelligence bill addressing the use of automated decision systems in hiring and other consequential matters failed to advance in the state assembly during the final hours of the 2025 legislative session on Friday.
Overview of the Bill
The bill, known as AB 1018, aimed to require companies and government agencies to notify individuals when automated decision systems were used for “consequential decisions” such as employment, housing, health care, and financial services. The legislation was authored by Democratic assemblymember Rebecca Bauer-Kahan.
Current Status and Future Plans
Assemblymember Bauer-Kahan announced a pause on voting until the next year to allow for “additional stakeholder engagement and productive conversations with the Governor’s office.” In her statement, she emphasized that this pause reflects a commitment to refining the critical legislation rather than a retreat from their responsibility to protect Californians. “We remain committed to advancing thoughtful protections against algorithmic discrimination,” she remarked.
Opposition to the Bill
The Business Software Alliance, a global trade association representing large technology companies, led an opposition campaign against the bill. They argued that the legislation would subject companies using AI systems to an “untested audit regime” that could deter the responsible adoption of AI tools throughout the state.
Craig Albright, senior vice president at BSA, stated, “Setting clear, workable, and consistent expectations for high-risk uses of AI ultimately furthers the adoption of technology and more widely spreads its benefits.” He highlighted the need for obligations based on different roles within the AI value chain and to ensure that low-risk uses of AI are not caught in a vague regulatory framework.
Amendments and Refinements
Since its introduction in February, AB 1018 underwent amendments aimed at narrowing the requirements for AI audits, clarifying which systems and “high-stakes” decisions are covered, and exempting low-risk tools such as spam filters. The amendments also added protections for trade secrets while limiting public disclosure of audit details. Furthermore, the bill refined how lawsuits and appeals would function and aligned itself more closely with existing civil rights laws.
Context of AI Regulation in Other States
The failure of AB 1018 comes shortly after the Colorado state legislature voted to delay the implementation of the Colorado AI Act, which also aims to regulate high-risk AI systems in sectors like hiring, lending, housing, insurance, and government services. This delay pushes the law’s effective date to the end of June next year, five months later than originally planned.
Similar to California’s initiative, Colorado’s legislation reflects the growing concern over the implications of AI technologies in critical decision-making processes.
