Navigating AI’s regulatory challenges

The increasing use of artificial intelligence, predictive analytics and other algorithm-based tools in the financial industry has led to some new steps in the ongoing dance of regulators and regulated. As Mark Schoeff Jr. reported in late July, the Securities and Exchange Commission has approved releasing a 243-page proposal for public comment that would require brokerages and investment advisory firms to review their use of these technologies and determine whether the algorithms optimize the interests of the firm or its financial advisors over the interests of investors.

If the proposed rule were adopted, firms would have to “eliminate or neutralize” conflicts, according to an SEC fact sheet. They also would have to implement related policies, procedures and record keeping. The theme of early comments in the press from brokerage and advisory executives was that the rule would make it harder for individuals to invest.

In a 3-2 vote, Democratic appointees favored the rule and Republican appointees opposed, with Commissioner Hester Peirce saying Regulation Best Interest already requires the financial advice business to address conflicts and that the proposal was “overbearing and would discourage financial advisors from using advanced technology.” Chairman Gary Gensler countered, saying the public should know whether algorithms built into predictive analytics optimize for client interests or those of advisors and firms.

In its current state, artificial intelligence is at once a miracle — pointing to life-saving solutions, for example, by being able to analyze voluminous medical data more efficiently and thoroughly than humans — and a Frankenstein. AI’s creators aren’t fully sure how it comes up with answers or solutions and are baffled by its “hallucinations,” which are totally fictitious fabrications that the algorithms concoct.

Like other technologies in the past, AI is bound to reshape the operations of the financial advice business, as well as those of the securities, investment and banking sectors of the economy. The issue is how to proceed within the framework of businesses that are already heavily regulated.

Perhaps one way to think of the path forward is to remember the reason for regulation in the first place. In the case of the SEC specifically, the reason was to protect the buyers of securities — the investing public — from the potential for deception, and worse, on the part of the creators of those securities (corporate issuers) and their sellers, namely Wall Street.

The widespread use of artificial intelligence only supercharges the potential for possible investor harm. For that reason, the SEC’s proposal is a good first step. Now it’s time for the industry to weigh in with its comments and join the dance.