• Private parties and governmental agencies are bringing lawsuits (including class actions) and investigations against companies accused of overstating the artificial intelligence (AI) capabilities of their products and services in a practice known as “AI washing.”
• “AI washing” cases have been filed in federal courts in California, Delaware, New York, and New Jersey.
• A federal court in California recently dismissed an AI washing case, holding the company’s statements about its AI capabilities were not actionable because they were forward-looking or amounted to puffery. 

While most of the litigation involving artificial intelligence (AI) making headlines centers on claims of copyright infringement (covered in several prior client alerts), businesses increasingly face lawsuits — or investigations by federal agencies — contending that they are “overhyping” their AI capabilities, use of AI, or the technological or business impacts of AI. These emerging claims are becoming known as “AI washing.”

Private lawsuits raising these claims have been brought against a variety of businesses in federal courts across the United States, including California, Delaware, New York, and New Jersey. Since 2024, the U.S. Attorney’s Office for the Southern District of New York has brought indictments for two executives for securities and wire fraud based on allegedly false and misleading statements about their companies’ proprietary AI capabilities and revenues. The Securities and Exchange Commission (SEC) and Federal Trade Commission (FTC) have also become more engaged in this space, bringing civil cases against companies and executives for alleged AI washing. Notably, the FTC branded its new law enforcement sweep as “Operation AI Comply.”

A federal court in California recently shed light on some AI washing theories leveled against a software as a service (SaaS) company in a securities class action lawsuit. In that case, the putative class alleged the SaaS company knowingly made false and misleading statements to investors about the company’s products and AI capabilities, customer feedback about the AI-enabled products, competitiveness in the market, and financial projections related to AI features. The plaintiffs, in their second amended complaint (meaning they had two prior attempts to plead their claims), claimed the alleged misrepresentations and omissions led to artificially inflated prices of the SaaS company’s stock, and subsequent disclosures caused the stock price to drop.

The court dismissed the case, holding that the challenged statements — including “our customers now can run AI and ML workloads on GPU-enabled” products, “we already have 14 AI features available to our customers,” “we expect customers will increasingly turn to [the SaaS company] as they build machine learning models and AI into their applications,” and “we’re winning deals against [a competitor] because we have the most comprehensive platform” — were not actionable under Section 10(b) of the Securities Exchange Act of 1934 and SEC Rule 10b-5.

In the court’s view, the challenged statements were either “forward-looking statements,” such as those concerning the SaaS company’s projected financial implications, accompanied by “meaningful cautionary language,” or amounted to “puffery or corporate optimism.” For example, the court held that the SaaS company properly notified investors that its “forward-looking statements are based upon current expectations and involve risks and uncertainties,” and that “any of the assumptions could prove inaccurate or incorrect and, therefore, there can be no assurance that the results contemplated in the forward-looking statements will be realized.” The court agreed that non-forward-looking statements — such as “we feel good about our ability to compete” — were inactionable because they were too vague or general to “give rise to any particular impression” about the SaaS company’s AI capabilities, particularly when language such as “I think” or “we feel” was used.

The court further disagreed with the plaintiffs that the statements about AI capabilities were false or misleading. For example, the plaintiffs, relying on either former employees or confidential witnesses, alleged that the SaaS company’s “AI product line was not matching its publicly touted capabilities,” and that an executive allegedly “exaggerated the Company’s AI capabilities.” The court held those allegations were “too generic because they do not identify as single AI feature that was not ‘matching’ its claimed capabilities,” and that the plaintiffs failed to “explain how the features were failing to meet expectations.”

Finally, the court found that the plaintiffs failed to sufficiently plead the required mental state, or “scienter,” to state a claim. Neither the allegedly “suspicious” timing of executive departures nor the individual defendants’ (who were sued along with the SaaS company) access to “core operations information” established that they knew the challenged statements were false or misleading.

Takeaways

• Public statements about AI capabilities, revenues, or market shares should be factually accurate or clearly identified as forward-looking.
• Forward-looking statements should always be accompanied by meaningful cautionary language.
• Companies using AI technologies should anticipate challenges not only from private plaintiffs but also from governmental actors.
• Given the prevalence of AI usage, businesses of all sizes and industries may potentially face “AI washing” claims.