Vague, Broad Definitions Plague Bills to Regulate AI

[Link]As California legislators enter the home stretch of the 2025 legislative session, they are facing a handful of bills that seek to regulate the artificial intelligence industry by imposing sweeping restrictions governed by vague, broad definitions of still-emerging technology.

These proposals, if signed into law, would threaten California's ability to win the global AI race - a race that could generate tens of billions of dollars in tax revenue for critical state programs and services.

Two bills that will soon face votes in the state Senate and Assembly seek new oversight of AI but do so using language that raises more questions than answers.

Senate Bill 7 (McNerney; D-Pleasanton) would impose new requirements on employers who use an automated decision system (ADS) related to employees or contractors. These systems ensure both efficiency and consistency across important business practices - in this case, the use of ADS for various operations involving employees.

Unfortunately, SB 7 seeks to provide employees with sweeping access to ADS data. The bill gives an employee or contractor the right to "correct" any data but is silent on how this would work or who would settle disputes.

"For example, a worker could go in and correct all time entries stating they clocked in late," wrote a coalition of business groups on Sept. 2.

SB 7 broadly defines ADS as any "employment-related decision" no matter how minor. And it pairs with these broad regulations a new private right of action, allowing almost any individual to seek enforcement through the courts.

Assembly Bill 1064 (Bauer-Kahan; D-Orinda) is a proposal promising to tackle deeply important and impactful issues pertaining to AI and California's children, with its supporters focusing primarily on the use of "chatbots."

But counterintuitively, the promise of AB 1064 - enhanced online protections for children - is undermined by bill language that would ban AI companies from developing the safe, secure, and effective AI technology that's broadly supported.

CalChamber and a coalition of business groups have pointed out that AB 1064 offers a very broad definition of a "covered product" that is trained by data related to someone under the age of 18.

These mandates stretch far beyond what would be considered regulation and would not only apply to widely used tools found on computers and smart phones but also technologies used in health care and education. It would even cover AI products designed to capture and block content that's harmful to children and teenagers.

"Restrictions in California this severe will disadvantage California companies training and developing AI technology in the state," wrote advocates for CalChamber and other opponents in a recent alert to legislators and staff.

Both bills, in essence, seek broad limits to solve more targeted problems - the kind of gap between intent and action that business advocates say would strike a heavy blow against California's hopes to maintain its role as the global leader in developing AI technology and products.

See the path to balanced AI regulation.