SEC Launches ‘AI Washing’ Enforcement

The Securities and Exchange Commission sanctioned two investment firms Monday for making false statements about how they use artificial intelligence in their business operations — the first enforcement actions we’ve seen under a crackdown on “AI washing” that SEC officials had been promising for several weeks. 

The two firms are Global Predictions Inc., an investment advisory firm based in San Francisco that runs; and Delphia USA, the U.S. subsidiary of an investment advisory firm in Toronto. Both businesses settled civil charges that they made misleading statements about how they use artificial intelligence to make investment recommendations to customers. 

As usual, the firms neither admitted nor denied the allegations against them, but both were censured and agreed to cease-and-desist orders. Delphia will pay a civil penalty of $225,000, and Global Predictions a civil penalty of $175,000.

So why should compliance officers care about two small-time investment firms busted for making false claims about how they use AI? Because these are the first two instances of the SEC cracking down on AI washing — and as sure as hallucinations flow from ChatGPT, more enforcement will come soon.

“As more and more investors consider using AI tools in making their investment decisions or deciding to invest in companies claiming to harness its transformational power, we are committed to protecting them against those engaged in AI washing,” said Gurbir Grewal, director of the SEC’s Division of Enforcement. “As today’s enforcement actions make clear to the investment industry: if you claim to use AI in your investment processes, you need to ensure that your representations are not false or misleading. And public issuers making claims about their AI adoption must also remain vigilant about similar misstatements that may be material to individuals’ investing decisions.”

The First Two Cases

For most corporate compliance officers, working outside the investment adviser industry, these two enforcement actions will feel a bit exotic because you’re not subject to the same rules that they are. Let’s quickly run through them anyway.

Delphia started a robo-adviser business (automated investment advice and data, somewhat like e-Trade or similar online trading platforms) in 2019, and planned to use artificial intelligence and machine learning to make investment recommendations to customers. In theory, Delphia would collect data from its clients (such as from social media, banking, credit card, online purchases, and so forth.) to use as inputs into its algorithms, along with externally sourced data as well. 

The AI would then do its thing and spit out recommendations. Delphia expressly said in its solicitations to clients that it “put[s] collective data to work to make our artificial intelligence smarter, so it can predict which companies and trends are about to make it big and invest in them before everyone else.”

Except, as described in the settlement order with Delphia, none of this AI stuff ever happened. The company had no such AI capabilities. Delphia even admitted as much to SEC examiners in 2021, but “continued to make certain false and misleading statements in advertisements regarding the use of client data in various formats through August 2023.” 

Global Predictions, meanwhile, went live last year and billed itself on its website as the “first regulated AI financial advisor” and made that same claim in emails and other messaging to prospective clients. Alas, as the SEC said in its settlement order, Global Predictions “could not produce documents to substantiate this claim.”  

Indeed, most of the charges against Global Predictions were compliance offenses that had nothing to do with AI, such as changing the terms of its advisory contract with clients and failing to implement certain procedures outlined in its compliance manual. 

AI Washing Enforcement Down the Road

I’m more interested in what enforcement chief Grewal said about public companies as a whole: “Public issuers making claims about their AI adoption must also remain vigilant about similar misstatements that may be material to individuals’ investing decisions.” 

That part could apply to any company in any industry. Moreover, one can imagine a wide range of circumstances where your disclosures about AI later get you into hot water with the SEC. 

For example, I look back at the SEC’s lawsuit against Brazilian mining giant Vale in 2022. The SEC sued Vale for making misleading statements to investors about its sustainability practices. In the 2010s, Vale proclaimed in corporate sustainability reports that it embraced the “strictest international practices” for dam safety. In fact, behind the scenes, Vale was falsifying safety audits and obtaining bogus safety certificates. Then one of its dams collapsed in southern Brazil, killing 272 people.

Vale and its executives ultimately faced severe criminal prosecution for that misconduct too — but the SEC claimed its own piece of Vale’s hide based on the company making false assurances about how it managed sustainability risk. So could other companies make similar rosy claims about their governance of AI risk, suffer some AI-related mishap, and expose themselves to SEC enforcement? That’s the AI-washing enforcement risk most companies would face. 

For example, most SEC enforcement over greenwashing (that is, misleading claims about your company’s ESG performance) has happened against financial or investment advisory firms — overselling how they consider ESG factors in investment strategy, making false statements about when they invest in green-energy businesses, and so forth. But beyond those sectors, greenwashing enforcement is (1) more rare; and (2) more often about a company’s failure to manage the ESG risk. That is, it looks something like the Vale case.

So imagine that your company says it uses artificial intelligence to, say, deliver the best possible services to its clients. Then, because your IT team bungled the data feeding into the AI, the AI ends up discriminating against certain types of consumers. Justice Department and Federal Trade Commission enforcement follows. 

Could the SEC then tag along with its own securities lawsuit, claiming that you didn’t have effective disclosure processes to warn investors that, oops, actually, your governance of AI was a mess?

I suspect we’ll find out soon enough.

Leave a Comment

You must be logged in to post a comment.