A proposal that would have exposed rental housing providers and other businesses to substantial penalties involving algorithm-assisted pricing will not advance this year.

Algorithm-assisted pricing typically uses automated systems — often powered by artificial intelligence (AI), or machine learning — to set, recommend, or adjust prices based on real-time data and predefined business rules.

Sen. Melissa Hurtado, D-Bakersfield, has decided not to move her bill — SB 295. The California Apartment Association opposed the bill and renewed its objections this year after engaging with the author’s office earlier in the legislative process.

The bill would prohibit specified uses of pricing algorithms that process confidential, nonpublic and competitively sensitive information from two or more competitors in the same market. Its definition of a commercial term expressly includes a rental rate, lease term, or occupancy level.

Under the bill, a housing provider could be subject to liability for using an algorithm’s recommendation if the provider knows or should know that the algorithm incorporates competitor data and that a competitor uses the algorithm or recommendation to set a similar commercial price in the same market.

CAA and its coalition partners, in a letter opposing the bill, argued that existing antitrust laws already prohibit price fixing, whether it occurs through people or technology, and warned that the proposal’s broad language could blur the distinction between unlawful collusion and legitimate use of pricing tools.

The coalition also warned that the measure could create uncertainty and increase costs for businesses using algorithms for lawful purposes.

The proposal would have allowed the California attorney general, district attorneys, county counsel and city attorneys to seek restitution, punitive damages, injunctions, attorney fees and civil penalties of up to $25,000 per violation.

For a business using an algorithmic recommendation, each month of use would have counted as a separate violation. For companies distributing algorithms or making recommendations, each authorized user or individual recommendation could be treated as a separate violation.