The rise of 'AI-washing' claims in securities class actions

In the race to tout "cutting-edge" AI capabilities, corporate optimism can outpace reality--and investors have taken note. In 2024 and 2025, shareholder plaintiffs brought approximately 29 federal securities class actions alleging "AI washing"--the overstatement or misrepresentation of a company's AI use, novelty or performance.

Several recent cases highlight the importance of exercising caution in making public disclosures. For example, in Black v. Snap Inc., plaintiffs alleged that Snap misled investors about the adoption of an alternative advertisers tracking tool and its effect on advertising revenue. No. 2:21-cv-08892 (C.D. Cal. filed Nov. 11, 2021). The parties have since agreed to a $65 million settlement, currently pending approval. Similarly, in Hoare v. Oddity Tech Ltd., plaintiffs alleged that the IPO offering documents overstated the company's AI capabilities and AI's contribution to sales. No. 1:24-cv-05037 (E.D.N.Y. filed July 19, 2024), transferred to 1:24-cv-06571 (S.D.N.Y. 2024).

Companies must also disclose risks posed by third-party AI tools. In Tamraz v. Reddit, Inc., investors alleged Reddit did not adequately disclose the risk Google's AI capabilities posed to Reddit's revenue. No. 3:25-cv-05144 (N.D. Cal. filed June 18, 2025). The complaint alleges that a significant portion of Reddit user traffic originates with a Google search and Google's AI overview reduces the need to click to Reddit for an answer, the complaint claims this feature diverted user traffic--and with it, advertisement revenue--away from Reddit.

While several AI-washing cases remain in early stages, others offer instructive lessons at the pleading stage. In Lamontagne v. Tesla, plaintiffs alleged that Tesla made materially misleading statements about the timeline, safety and capability of its autonomous driving technology. No. 23-cv-00869, 2024 WL 4353010 (N.D. Cal. Mar. 28, 2025). Ultimately, the court granted Tesla's motion to dismiss, finding that many of the challenged statements were protected by the safe harbor because they were forward-looking, constituted mere puffery or contained adequate cautionary language. For example, the court found that statements regarding projected launch dates were plainly forward-looking statements, and statements that the technology is "super," "superhuman," and that the company wants to "get as close to perfection as possible" were held to be vague statements of corporate optimism that are not actionable. Although the case was dismissed, statements that safety "bears out in the statistics" and references to "objective numbers" regarding the low likelihood of injury were found to be actionable.

A similar lesson, albeit with a different outcome, is highlighted by In re GigaCloud Technology Inc. Securities Litigation. No. 23-cv-10645, 2025 WL 307378 (S.D.N.Y. Jan. 27, 2025). Plaintiffs' claims that GigaCloud misrepresented the sophistication of its technology were held to be insufficient at the motion to dismiss stage because "vague corporate-speak" describing the strength of technology is mere puffery. However, plaintiffs' specific allegations that GigaCloud's claimed use of AI in its logistics operations was false were adequately pled. Accordingly, while the alleged misstatements regarding GigaCloud's marketplace activities were dismissed, the alleged misstatements regarding GigaCloud use of complex AI software to optimize logistics survived the motion to dismiss. In support of the plaintiffs' argument that GigaCloud did not use AI in its logistics, plaintiffs detailed that GigaCloud's logistics system involved manual computations and at least 100 IT employees. Plaintiffs' claims were supported by information from nine former employees who directly worked on the company's logistics systems. The court concluded that plaintiffs' claims rested on specific factual statements about GigaCloud's use of AI--alleged to be false and supported by specific contradictory information--and that those allegations met the motion to dismiss standard.

The trend of AI-related securities suits will continue to test the balance between rapid innovation and cautious disclosure. Companies can reduce their exposure by:  

• Framing descriptions of in‐development AI as forward-looking, employing safe-harbor disclaimers, and distinguishing between exploratory and commercial applications.

• Limiting statements about current technology use to those well-supported by current data--vague or grandiose claims can become the basis for fraud allegations if outcomes disappoint.

• Proactively assessing competitive technology and disclosing material uncertainties.

• Acting swiftly when litigation arises by marshaling data, engaging experts and pressure-testing the complaint while maintaining consistent and measured communications.

By staying proactive, companies can effectively navigate the challenges of securities litigation in this dynamic landscape.