|
Getting your Trinity Audio player ready...
|
A piece of proposed legislation from Assemblymember Rebecca Bauer-Kahan (D-Orinda) seeking to eliminate AI bias has advanced in the State Legislature in the latest move on multiple measures from the Tri-Valley representative aimed at regulating the technology.
Assembly Bill 2930 would seek to prevent bias in AI-powered Automated Decision Tools, widely used for credit decisions and increasingly in other domains such as employment screening, insurance eligibility, and healthcare decisions.
“Bias is embedded in the decisions made about our lives,” Bauer-Kahan said in an announcement on April 16. “The difference with AI is that it’s completely opaque. Without regulation, these tools can reinforce inequality without our knowledge. AB 2930 continues our work to ensure that AI fulfills its promises of progress and does not drag us backwards with biased results.”
The bill advanced through its first hearing at the Assembly Committee on Privacy and Consumer Protection on April 16, where it was moved forward with a vote of 8 to 3. It features the support of two major tech companies — Microsoft and Pleasanton-based Workday.
“With a growing appetite to regulate AI at the state level, California can lead the nation in establishing meaningful guardrails,” said Chandler Morse, vice president of public policy at Workday. “This new bill would establish concrete requirements and strong protections for consequential decisions driven by AI.”
While the advancement served as a milestone for Bauer-Kahan’s office and the bill’s supporters, this month’s vote was just a first step in the review process at the state legislature before the bill makes its way to the governor’s desk. From here, it is headed to the Appropriations Committee, then to the Assembly floor, then the State Senate if it advances through the next two steps.
The bill was first introduced in February along with other proposed legislation that would seek to regulate AI, including establishing a definition for the term, creating a registry for AI tools, and restricting digital duplication of deceased performers.
The latter, AB 1836, would prevent the use of generative AI for replicating images or sounds of a performer’s face, voice, or likeness following their death, boasting a sponsorship by SAG-AFTRA.
“Over 30 years ago the California legislature established a property right in voice and likeness for deceased individuals,” said Duncan Crabtree-Ireland, chief negotiator and national executive director for SAG-AFTRA. “That law, written decades ago, included express exemptions for audiovisual works and television programs allowing producers of that content to use performers’ voice and likeness even without the consent otherwise required by the law.”
Crabtree-Ireland noted that the legislation already in the books should be updated to account for advances in AI and new challenges in the present day.
“Nobody at that time imagined that artificial intelligence would arrive and make it possible to easily create new performances by deceased performers,” Crabtree-Ireland said. “Nobody thinks it is acceptable to replicate deceased performers without the consent of their family. Families are offended by this and it’s bad public policy for California.”
Bauer-Kahan noted that generative AI is already being used to take advantage of artists, pointing to the urgency of the proposed legislation.
“Deceased performers could never have guessed their images would be re-animated through technological means,” Bauer-Kahan said. “We must ensure that an artist’s intellectual property is not taken advantage of by the unanticipated capacities of AI to make new versions of an artist’s work.”
She added that the package of bills is the first step toward developing a legislative system that would regulate the rapidly changing enterprise of AI and seek to prevent potential harms in the present and future.
“Creating a regulatory framework where AI tools are tracked, understood, and clearly defined is essential to protect Californians,” Bauer-Kahan said.