N'achetez pas les mots à la mode : Le "lavage par l'IA" a du plomb dans l'aile

Since the release of ChatGPT 3.5 in November 2022, public interest in artificial intelligence (AI) has surged in a classic example of a hype cycle. As with past technological breakthroughs, companies may be tempted to overstate their AI capabilities to draw investor attention.

But that may be coming to a swift end as the U.S. Securities and Exchange Commission (SEC) has begun paying close attention to this “AI Washing” trend and warning organizations against overstatement.[1]

What is AI Washing?

“AI Washing” is the intentional overstating of a product or service’s AI capabilities to make such product or service appear more innovative or intelligent than it actually is, and thus “artificially” inflating sales or engagement. The phrase stems from “greenwashing” (which itself came from “whitewashing”), a term frequently used to describe companies, products, or services that exaggerate their efforts to reduce environmental impact for the sake of appealing to environmentally conscious consumers.

Regulators have been warning about the risks of AI Washing for some time. SEC Chair Gary Gensler, while speaking at an AI conference in December 2023, cautioned: “Don’t do it…. One shouldn’t greenwash, and one shouldn’t AI wash. I don’t know how else to say it.” Reiterating those sentiments in prepared remarks at Yale Law School in February 2024, Gensler again cautioned companies against overstating their AI capabilities: “If a company is raising money from the public, though, it needs to be truthful about its use of AI and associated risk…. As AI disclosures by SEC registrants increase, the basics of good securities lawyering still apply. Claims about prospects should have a reasonable basis, and investors should be told that basis.”

More recently, on April 15, 2024, Gurbir Grewal, Director of the SEC’s Division of Enforcement, warned:

If you are rushing to make claims about using AI in your investment processes to capitalize on growing investor interest, stop. Take a step back, and ask yourselves: do these representations accurately reflect what we are doing or are they simply aspirational? If it’s the latter, your actions may constitute the type of “AI-washing” that violates the federal securities laws.

Yet, the impact of AI on our lives will continue to expand, and how AI is disclosed and discussed by companies and firms will continue to evolve in tandem with the risks associated with such disclosures. As Mark Zuckerberg posited during Meta’s July 2024 earnings call, AI is going to affect almost every companies’ products in some way; specifying that “this is why there are all the jokes about how all the tech CEOs get on these earnings calls and just talk about AI the whole time.” This remark highlights the pressure that CEOs and companies face to hop on and keep up with the AI bandwagon.

Recent SEC Enforcement Actions

On March 18, 2024, the SEC announced its first ever settled charges against two investment advisers, Delphia and Global Predictions, for violating antifraud provisions of the Investment Advisors Act of 1940 through purported misrepresentations about their use of AI. Both companies claimed that they utilized certain AI technologies to attract investors, but did not actually use those AI capabilities.

Delphia publicly claimed that it used AI and machine learning to analyze client data to inform investment decisions, purporting that it “put[s] collective data to work to make [its] artificial intelligence smarter so it can predict which companies and trends are about to make it big and invest in them before everyone else.” According to the SEC, Delphia’s claims were false and misleading because Delphia did not have the AI capabilities it publicly represented.

Similarly, the SEC alleged that Global Predictions made false and misleading statements about its AI expertise as the “first regulated AI financial advisor” and its technologies that incorporated “[e]xpert AI-driven forecasts.”

Delphia and Global Predictions both settled violations of Section 206(2) of the Advisers Act for their false and misleading statements. Both companies were also found to have violated the Marketing Rule, which makes it unlawful for registered investment advisers to produce advertisements that include any untrue statement of material fact. Delphia paid a civil penalty of US$225,000 and Global Predictions paid a civil penalty of US$175,000.

More recently, on June 11, 2024, the SEC announced litigated charges against the CEO and founder of a now-shuttered AI recruitment startup for alleged violations of the antifraud provisions of the Securities Act of 1933 and Securities Exchange Act of 1934. In its complaint, the SEC alleged that the CEO “engaged in old school fraud using new school buzzwords like “artificial intelligence” and “automation.” The SEC further alleged that the CEO defrauded investors of at least US$21 million by making misleading statements about the quantity and quality of the company’s customers, the number of candidates in its platform, and the company’s revenue.

Director Grewal concluded his remarks in the press release with this admonition, “As more and more people seek out AI-related investment opportunities, we will continue to police the markets against AI-washing and the type of misconduct alleged in today’s complaint. But at the same time, it is critical for investors to beware of companies exploiting the fanfare around artificial intelligence to raise funds.”

Principaux enseignements

As evidenced by these enforcement actions, the SEC is taking AI Washing very seriously and companies should be diligent and honest to ensure that they do not engage in this practice – either intentionally or inadvertently. To ensure compliance with the SEC’s protocols, companies should consider the following:

• Fully and accurately disclose your AI usage. AI capabilities vary, so avoid using boilerplate language that is either overly vague or broad. Also avoid using vague or exaggerated claims and hypothetical examples to describe what your AI model is capable of doing.
• Be specific about the nature and extent of your AI technologies, the role AI plays in your business operations, and any potential risks or limitations associated with AI.
• Understand how your key service providers employ and use AI, as that will likely be the focus of future SEC oversight rules.
• Provide details about the company’s AI implementations, including which processes or products it impacts, the extent of its deployment, and any measurable outcomes.
• Establish and implement an AI governance framework to provide “scaffolding” for AI initiatives and ensure they align with the company’s goals and ethical standards.
• Provide training for company marketing teams to ensure that technologies are properly labeled as “AI.” Many technologies and algorithms do not actually qualify as AI but may be easily mistaken for it, so being aware of this before creating marketing materials will be crucial to avoid inadvertent AI Washing.
• Require that all public statements or advertising produced by the company regarding AI technologies be reviewed by the company’s legal team to ensure the accuracy of such statements.
• Monitor the company’s use and evolution of AI technologies, as well as external public statements regarding the company’s use of AI technology and correct any misstatements or inaccuracies that may arise.
• Regularly update shareholders and other stakeholders on the progress, changes, and improvements in AI initiatives.

Special thanks to Natalie Smith, a summer associate in Foley’s New York office, for her contributions to this article.

[1] In addition to the risks of SEC enforcement, companies also face the threat of private securities class actions. Cornerstone Research found an uptick in securities class actions with allegations of AI-related deceptions. According to Cornerstone, investors filed six AI-related class actions between January and June of 2024, compared to six such actions in all of 2023.