Partner James T. Christie and Of Counsel Nicholas D. Manningham are authors of an informative article on “The Rising Tide of AI-Washing Cases in Securities Litigation,” published by Corporate Compliance Insights.

In the article, James and Nick examine the increasing prevalence of ‘AI-washing,’ a practice where companies exaggerate or misrepresent a company’s AI capabilities, or the role that it plays in their products or services.  The authors assert that, as AI transforms financial markets, overstating claims of its capability to enhance business operations can mislead investors, exposing companies to securities fraud cases.

James and Nick point to two cases as examples of companies that engaged in AI-washing and exposed themselves to securities fraud liability.  In In re Opendoor Technologies Incorporated Securities Litigation, the company touted a competitive advantage with its AI-powered pricing algorithm but was unable to prove that the algorithm performed the revolutionary adjustments they claimed without relying on a human-driven process.  In In re Upstart Holdings, Inc. Securities Litigation, the court found the company made actionable misstatements about its AI underwriting model’s advantageous capabilities, as the AI model did not in fact provide those advantages.  Both cases illustrate the growing issue of companies risking securities fraud liability by over-exaggerating or misrepresenting their AI capabilities.

James and Nick specify, “if a company claims that adding AI miraculously solves a previously unsolvable problem, and it sounds too good to be true, it probably is.”  To emphasize the importance of companies being transparent regarding the strengths and limitations of their AI capabilities, they also note that “where AI plays a material role in core business functions, companies should also disclose known limitations, assumptions and circumstances under which the technology may perform less effectively, particularly during periods of market stress or atypical conditions.”  

Clear and reasonably specific descriptions of AI systems, using precise language to avoid vague or exaggerated claims and providing evidence to support their claims, should become the minimum for what transparent AI reporting looks like.  James and Nick conclude, “As this technology becomes integral to corporate strategy and investor expectations, plaintiffs and regulators alike are scrutinizing the accuracy and completeness of AI-related disclosures.”