Advisors love AI, but most are scared of the compliance risks

Advisors are often put off by new rules and regulations. But when it comes to generative AI and similar technologies, demand is growing for some sort of guidance from industry watchdogs.

Robert Cruz, the vice president of regulatory and information governance at the compliance consultant Smarsh, said financial advisors have general trepidation about putting artificial intelligence, machine learning and other advances to use in wealth management. The fear is largely born out of a desire to avoid tripping long-standing regulations that may or may not pertain to AI.

"The problem is that you have a lot of rules that are antiquated, that were written many years ago," Cruz said. "And we still have these nebulous concepts like 'business as such.' How do you apply that to the use of Gen AI?"

'Regulatory, compliance concerns' top reasons to not embrace AI

Of the various reasons for not adopting AI cited by respondents to the FP AI Readiness Survey, the biggest was "regulatory and compliance concerns." That answer, given by 44% of the poll takers, exceeded "data security concerns" (40%), "concerns about uncontrolled tech and mistakes" (36%), and "ethical concerns/potential for bias" (33%.)

Cruz said there is certainly room for regulators at the Securities and Exchange Commission and other agencies to shed a bit more light on the ways advisors can use AI without running compliance risks. But the chances they will do so any time soon are slim.

"It's very unlikely we're going to see any significant regulation of AI at the federal level, in particular from the SEC," Cruz said. "I think what the SEC is saying under the new chairman is that they're trying to simplify the burden of regulatory compliance and trying to eliminate more rules than they create."

The SEC in 2023, under its former chairman, Gary Gensler, put forward a sweeping proposal to regulate advisors' use of AI and similar technologies in "predictive analytics." The proposed rule would have made advisors explicitly responsible for finding and then negating any conflicts of interest that could be created by their use of AI, machine learning and other systems.

The proposal met with strong resistance from lobbying groups and other industry representatives. Many of the critics argued its prohibitions could apply to simple spreadsheets and other applications long in use.

The SEC's AI proposal was one of several pending rules that were dropped in June as part of current Chairman Paul Atkins' push to take a lighter regulatory approach to securities markets. The SEC hasn't since put forward a replacement.

Existing rules apply to AI, but it's not always clear how

Republican commissioners like Mark Uyeda and Hester Peirce, who hold sway in the current administration, have generally favored using existing rules rather than new ones to prevent abuses of AI and other new technologies. Firms, for instance, are already barred from making misleading statements by the SEC's marketing rule. The same prohibitions apply to misinformation produced by generative AI.

"In the SEC's view, rules are meant to be technology agnostic," Cruz said. "If there's a financial crime, there's a financial crime, regardless of whether it's delivered via AI, a text message or a carrier pigeon."

And the SEC has proved willing to bring enforcement actions under the existing rules. In March 2024, it reached a $400,000 settlement with two firms accused of making unsubstantiated claims about their use of AI in providing investment recommendations. Before that announcement, Gensler had warned the industry about so-called "AI washing" — a term referring to exaggerated claims about artificial intelligence, much as "green washing" refers to unsubstantiated statements about environmental benefits.

Similarly, advisors and brokers are required by industry conduct standards like the fiduciary duty or Regulation Best Interest to either eliminate or disclose conflicts of interest. To regulators, it makes little difference if a conflict came from AI; it still has to be negated or at least mitigated.

Ben Marzouk, a partner and securities regulation lawyer at Eversheds Sutherland's Washington, D.C., offices, said he could see the current SEC introducing a rule that would make the fiduciary conduct standard and Regulation Best Interest explicitly apply to AI uses.

"That would be the more narrow path to include artificial intelligence within their regulations without a whole new AI rule," he said. "But I don't think they even have much appetite to do that."

Marzouk said a lack of specific regulation could ultimately hinder wealth managers' willingness to try out new uses for AI. 

"It's certainly a hot topic among the regulators, and it's going to get attention just because it's a new and shiny thing," Marzouk said. "If you were to use it in a customer-facing way, and it explodes on you, I think that's a really easy case for a regulator to bring." 

Cruz agreed that wealth managers would welcome some sort of guidance on their use of AI, while acknowledging the difficulties in drawing up the sought-for regulations.

"I think when you make a rule that's technology-specific, the problem is, by the time you create the rule, the technology has moved on," he said.

What firms are using AI for, and what they're not

Even amid the regulatory uncertainty, wealth managers are discovering many uses for AI. The Financial Industry Regulatory Authority, the broker-dealer industry's self-regulator, has noted instances of firms using AI for everything from responding to routine client inquiries and reviewing and classifying incoming email, to conducting market research and meeting various regulatory requirements.

In their latest earnings calls, executives at big Wall Street players like Morgan Stanley, Merrill and Wells Fargo spoke of how they're investing even more in AI. 

"These include the DevGen.AI tool, which enhances developer efficiency by modernizing code; Parable, an interactive tool that quickly analyzes and summarizes data; and LeadIQ, our AI-powered lead distribution platform, focusing on matching workplace and self-directed relationships and facilitating engagement with our financial advisors," Morgan Stanley Chief Financial Officer Sharon Yeshaya told analysts last week. "Together, these use cases are laying the foundation to drive productivity across the firm."

Financial Planning's recent industry survey revealed both enthusiasm for new uses for AI and misgivings about possible legal and other complications. Asked how much they agreed with the statement "AI is making me more effective," 63% of the respondents said they "totally" or "mostly" agree. But then presented with the statement that "risk management and legal considerations make it harder to adopt new technologies," nearly the same percentage also expressed agreement.

Leslie Norman, the chief technology officer at Dynasty Financial Partners, said in an interview that advisors with her firm have been rushing to make use of AI. For many, the biggest help from the technology is in its ability to shave time off an often-routine task like preparing for a client meeting.

What once could take hours, as advisors had to search through old emails and notes on previous meetings to remember topics discussed and priorities set, can now be done in a matter of minutes, Norman said.

"It kind of smacks you in the face when you see client meeting prep go from being a laborious task to something that happens in a second," Norman said. "AI can tee up a note for you about referencing an interaction that you had with the client three years ago that's embedded in an email exchange.

Like many other firms in wealth management, Dynasty is hesitant to have AI make actual investment recommendations. That's because investing decisions and judgement calls often are contingent on an understanding of human psychology.

"As an advisor, you have to ask yourself if it makes sense, based on the person that you're talking to, to make a recommendation for switching their portfolio from a growth strategy to preservation mode, right?" Norman said. "That is the type of deeply personal decision that I think a human is best positioned to make."  

Norman said when Dynasty does field questions from advisors about AI compliance, it's often because they want to make sure their use of it doesn't risk exposing private client data. That's in part why Dynasty Financial Partners, which provides investing, marketing, compliance and other types of services to more than 500 affiliated independent advisors, has built up its technological expertise.

"What we tell people is this is not a space that you should be navigating alone," Norman said. "Now with AI, it adds a whole other layer of complication that you need to have some expertise in order to figure out."