An artificial intelligence regulatory bill has now cleared a vote in the California state Senate’s Judiciary panel after passing the Assembly, despite concerns from the Business Software Alliance that “rigid mandates” in areas such as risk assessments could undermine AI innovation.
"BSA urges California to reverse course on AB 1018. Recent amendments did not change that the bill remains a sweeping and flawed attempt to regulate artificial intelligence (AI) that misunderstands how AI is built and deployed,” senior vice president of U.S. government relations Craig Albright said in a July 16 statement.
The legislation, AB 1018 by Assemblywoman Rebecca Bauer-Kahan (D), passed the Assembly in June.
AB 1018 seeks to “create a comprehensive transparency regime for developers and deployers of ADS that are used for consequential decision-making,” according to an Assembly committee analysis. “The author’s goal is to prevent algorithmic discrimination on the basis of protected characteristics and to ensure that ADS can be trusted to perform reliably and accurately.”
But BSA’s Albright said, “AB 1018 imposes rigid mandates on a broad range of businesses that develop and create tools with AI. It unwisely requires them to conduct performance evaluations and risk assessments for downstream uses they have no visibility into or control over. AB 1018 makes a large part of the AI value chain responsible for deployment choices to which they are not a party, including AI fine-tuning and use cases.”
“The result,” Albright said, “is unworkable regulation that creates more confusion than clarity and ultimately harms AI development and adoption.”
He argued, “Effective AI regulation should assign responsibility based on real-world roles and risks. Both developers and deployers must play a part -- but AB 1018 gets it wrong. … [Recent] amendments to AB 1018 are not sufficient; California lawmakers should stop AB 1018 and revisit its provisions with a better understanding of how the AI supply chain works in practice."
California Gov. Gavin Newsom (D) says the state’s AI policy should track with the recommendations of an expert panel that he appointed after vetoing a high-profile AI bill last year.