In recent months, the SEC Chair and other government officials have made it clear that they are closely watching for companies that try to catch the AI wave by making exaggerated or false claims about the AI capabilities of their products or services. In the latest example of the crack down on these kinds of AI-related misrepresentations, both the U.S. Attorney  for the Southern District of New York and the SEC have filed charges against the former CEO and Chairman of the tech company Kubient. The government alleges that the executive, Paul Roberts, inflated the company’s revenues and also lied about one of its signature products, an AI-powered tool that was supposed to detect ad fraud in the digital advertising industry. There are some interesting features of the charges against Roberts, as discussed below.

 The Southern District of New York’s U.S. Attorney’s Office’s September 16, 2024, press release about the criminal charges against Roberts, including a link to the criminal information filed against him, can be found here. The SEC’s September 16, 2024, press release about the charges it filed against Roberts and two other Kubient executives, including links to the agency’s civil complaints, can be found here.

Background and Allegations

Kubient is a digital advertising technology company. It completed an IPO in August 2020 and its shares traded on Nasdaq until November 2023. Kubient’s flagship product was a proprietary fraud detection tool, Kubient Artificial Intelligence (KAI).

The charges filed against Roberts allege that as part of “a deliberate scheme” to “concoct phony revenue” in order to deceive investors and the company’s auditor, Kubient and another unnamed entity entered several contracts in which they agreed to provide services to each other for $1.3 million each, with Kubient to provide its flagship AI fraud detection services to the other company. The money was paid under the services, and Kubient recognized the payments as revenue, but no services were actually rendered between the companies. Kubient reported the phony revenue in documents filed with the SEC in connection with its $12.5 million IPO and subsequent $20 million secondary offering.

The agencies’ charges also allege, as summarized in the U.S. Attorney Office’s press release, that Roberts “repeatedly made material misrepresentations in SEC filings about the efficacy of KAI in identifying and preventing digital ad fraud, including in connection with Kubient’s initial and secondary public offerings, when Kubient was touting KAI as one of the company’s premier products that would differentiate it from its competitors.” The government alleges that Roberts “made these statements knowing that they were false.” The government also alleges that Kubient’s IPO and its secondary public offering “would not have been possible without Robert’s fraudulent misrepresentations about KAI and the KAI transactions.”

The U.S. Attorney’s Office’s press release states that Roberts has pled guilty to one count of securities fraud. The SEC’s press release states that as a partial settlement Roberts has consented to injunctions from future violations of the charged violations while agreeing to litigate the appropriate remedies.

Discussion

In making their charges against Roberts, the government’s complaints, as well as the press releases about the complaints, do not use the term “AI washing.” One commonly accepted definition of the term is that AI washing is a marketing tactic where a company. Microsoft’s AI-powered product Co-Pilot provided me with the following definition of AI Washing: “a marketing tactic where a company or organization falsely claims or exaggerates the use of artificial intelligence (AI) in their products, services, or operations.” Using this AI-generated definition as I guide, I think it is fair to say that the government’s charges against Roberts amount to AI washing, even though the government itself did not use that specific term.

As I have noted in prior posts on this blog, the SEC has previously brought AI washing-related enforcement actions. In at least one prior instance, the SEC enforcement action, like the one here, was accompanied by a parallel criminal prosecution as well. As discussed here, the SEC charged the founder and CEO of an AI- powered employment recruitment company, Joonko Diversity, Inc., with making false claims about the company’s services. In addition to the SEC’s action against the founder and CEO, the U.S. Attorney’s Office for the Southern District of New York filed a criminal action against the individual as well.

But while there have been prior SEC enforcement actions and even criminal actions in which the government has alleged that companies and their executives engaged in securities fraud by AI washing, none of the prior cases of which I am aware have involved a publicly traded company. (Joonko was a privately-held company). As far as I am aware, the government’s charges against Roberts in connection with Kubient represent the first example of enforcement actions being brought against a listed public company based on AI washing-related allegations. (If any readers out there are aware of any prior action, please let me know.)

It might be argued that what went on here was basically what some commentators have referred to as “old school fraud.” But an important part of what went on here is the use of false AI-related claims in order to attract investor attention and perhaps to catch some of the AI lightning in order to maximize the returns on the fraudulent scheme.

It is clear that in the current environment, when investors are eager to capitalize on the new technological capabilities that AI promises, the government is worried that companies will seek to mislead investors by using AI-related buzzwords to attract attention. The government’s crackdown here underscores the risks that companies face when they misrepresent their AI-related capabilities – and the parallel criminal action further highlights the risks companies may face when they exaggerate of falsity their AI credentials. Of course, as readers of this site are well aware, the associated risks also include the possibility of AI-related securities class action litigation as well.