AI-Washing and Recent SEC Enforcement

The SEC is not taking lightly to investment managers bragging publicly about using AI in their investment process and then failing to do so. The regulator fined two investment advisors $400k for misleading claims that it considers to constitute 'AI washing' :

The SEC's Director of Enforcement is cited saying "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."

The Cases

Based off the press release, the merits of both cases seem rather straightforward.

Delphia is said to claim that it "use of AI and machine learning that incorporated client data in its investment process" amongst other things "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". Anyone that has seen SEC enforcement knows that using phrases such as "making it big" or "investing before everyone else" are red flags and almost surely attract attention from the regulator.

Global Predictions, on the other hand, apparently had a multitude of failings. One of which was that it claimed to be the "first regulated AI financial advisor". Again, using the phrase "regulated" to imply some sort of approval from the regulator is a red flag. Plus, according to the order there seems to be some other non-AI misrepresentations that the SEC took issue with as well.

The Implication

The bigger question though is what does this mean for the industry? Does the SEC want to police every claim of AI usage in investment management? Does it want to make rules what constitutes AI and what doesn't?

Governments around the world seem to see AI as a pinata for new regulation. Both the US and the EU have recently passed AI regulation. The EU even went so far and boasted about its achievement to create "the first comprehensive regulation on AI by a major regulator anywhere in the world". Leaving aside whether this is something that should be boasted about (it shouldn't!), rule-making for the industry has started.

So far though, most regulation is aimed toward the technology companies that build and train foundational models: the Googles, Amazons, and Microsofts of the world. It's about which data can be used, how privacy needs to be protected, etc. And it takes a fairly inclusive approach, on balance including more companies and technologies under the regulatory umbrella than excluding them.

The SEC's actions are different. They are aimed at AI applications, not the models. They implicitly open up the discussion of what constitutes AI in investment management. Is it ok for an investment advisor that claims to use AI to use a regression model? An old-school neural network? How many parameters does the model need to have to be considered AI?

Maybe these actions were just a shot across the bow. Maybe the SEC just wanted to make sure that the industry knows that it's watching. But maybe it's the start of a new era of regulation in investment management. And maybe it will finally answer whether a regression model is AI or not.