A Securities and Exchange Commission proposal designed to regulate the use of predictive data analytics and AI by broker-dealers and investment advisors could slow the sector's pace of innovation and increase costs for clients, panelists said at a Security Traders Association conference Friday.
The SEC's proposal, released in July, would require broker-dealers and investment advisors to "eliminate or neutralize" conflicts of interest related to the use of "covered" technologies in investor interactions, such as analytical functions, algorithms and models that predict, guide or direct investment-related behaviors. Firms would also be required to maintain policies and recordkeeping related to possible conflicts of interest.
Panelists, which included leaders at Robinhood and the American Securities Association, critiqued the proposal for being vague, onerous and costly to broker-dealers and investment advisors.
Matt Billings, vice president of brokerage and president at Robinhood Financial and Robinhood Securities, said the proposal feels rushed. He added any rulemaking should start with extensive cross-industry conversation, data analysis and economic analysis.
"All the costs that we discussed that we'll absorb, that's going to leech down to the investor," Billings said. "Then, with all this intense amount of policy and procedures and recordkeeping, maybe a firm decides not to offer that covered technology, or maybe they slow down their pace of innovation … the result of that is a worse experience for the client."
SEC Chairman Gary Gensler
"If an investment advisor, think about a robo advisor, is telling you their advice and it's purely based on your family and your wealth and so forth, great," Gensler said at the hearing. "But if they're also taking into account their own interest and profits, their revenues and the like, therein lies a potential conflict. And whether they're using machine learning or other data analytics, there may be a conflict there."
Many banks, such as Morgan Stanley, JPMorgan Chase, Goldman Sachs and BNY Mellon, would also be impacted by the rule. Morgan Stanley is using generative AI to assist financial advisors with sourcing information,
In comment letters to the SEC, JP Morgan and Morgan Stanley both said the proposal was unnecessary and overly burdensome. Panelists and comment letters also note that the proposed rule extends beyond artificial intelligence and emerging technologies to include common spreadsheets. JPMorgan said in its letter that the SEC should propose a rule "narrowly targeted to address" concerns around technology-driven interactions "where conflicts of interest are not disclosed or evident to a reasonable retail investor."
Financial institutions are generally familiar with ChatGPT, but other options exist such as language models sold by Anthropic and Cohere. Research firms, academics and others are racing to develop benchmarks for comparison shopping.
"In accordance with our core values, Morgan Stanley acknowledges the SEC's objectives to protect investors and support responsible implementation of technology, including technology such as predictive data analytics," Morgan Stanley wrote in its comment. "However, we respectfully believe the proposed rules' scope and significant obligations are overly broad in light of the robust and comprehensive regulatory framework that is already in effect for broker-dealers and investment advisers."
Elad Roisman, a partner at Cravath, Swaine & Moore, said on the Friday panel that he doesn't think the SEC adequately identified the problem it's solving, or issues that can't be addressed with existing regulation.
Brett Redfearn, an investment advisor who previously served in capital market leadership roles at JPMorgan Chase, the SEC and Coinbase, said in the panel that he thinks this rule goes beyond investor protection.
"I really do believe that we have to look a lot harder on what's happening in our market, and the way in which we use technology without taking an overly simplistic and probably not fully thought-out analysis that results in this kind of attack on tech," Redfearn said.
