California trains its sights on algorithmic collusion

The amendment addresses algorithmic pricing in two ways. First, it bans using or distributing a "common pricing algorithm"—which the statute defines as technology that two or more entities use to recommend prices or commercial terms based on competitor data—as part of an agreement to restrain trade under the Cartwright Act. Second, it makes it illegal to use or distribute such an algorithm while coercing someone else to follow the algorithm's recommendations for the same or similar products or services.

One important wrinkle: the legislative history makes clear that these provisions apply whether the algorithm uses private or public competitor data.² As a result, companies looking to use pricing algorithms need to scrutinize not just the data that users contribute, but also any data the algorithm provider pulls from public sources. While the statute creates liability for "coercing" another to adhere to recommended prices, it doesn't define coercion—leaving courts to sort out what counts.

The new law also clarifies what plaintiffs need to allege to bring a Cartwright Act claim. Plaintiffs must include "factual allegations demonstrating that the existence of a contract . . . is plausible" but don't need to "allege facts tending to exclude the possibility of independent action." Most California courts already applied this plausibility standard at the motion to dismiss stage, so this may just reinforce existing practice rather than break new ground.

California isn't the only state paying attention to algorithmic pricing, though it's taking a collusion-focused approach. Other states have passed or are considering laws that tackle this issue differently, some focusing on specific markets like housing. For example, New York recently banned landlords from using algorithmic software that analyzes other property managers’ information to set rental rates, lease terms, and occupancy levels.³ Connecticut passed a similar ban that specifically prohibits the use of competitors’ non-public data as an input for algorithmic rental rate recommendations.⁴ New York also now requires businesses to disclose when prices are set by “surveillance pricing”—a tool that analyzes consumers’ data to set personalized prices—with other states like Illinois and Pennsylvania considering similar measures.⁵

The new California law, combined with the proliferating patchwork of state regulations, adds another layer of complexity to decisions companies face when adopting AI. While federal courts are developing a consensus around how algorithmic pricing providers collect and analyze user data, these new state laws may shift the risk profile. We will continue to monitor these developments using our multidisciplinary team, which includes experienced practitioners in antitrust, privacy, IP, and technology transactions. A&O Shearman helps companies navigate this fast-moving landscape as they bring AI into their operations.