Advisers Are Skeptical of AI’s Potential. This Economist Says It May Be at Their Own Loss

Earlier this year, behavioral economist and retirement industry entrepreneur Shlomo Benartzi took to LinkedIn to ask his network of financial professionals a question: When will AI outperform humans at all tasks?

Benartzi, a UCLA professor emeritus and founder of a retirement planning tool called PensionPlus, has been putting his attention toward artificial intelligence for possible breakthroughs in providing retirement plan participants with holistic advice. He believes, if done correctly, there is potential to provide participants with financial guidance that can improve saving, budgeting and retirement income outcomes in ways current methods have failed to provide.

Shlomo Benartzi

“AI would allow us to virtually get financial advice in channels for discussion as if you were sitting with an adviser 24/7,” he says. “It’s the future.”

Benartzi was surprised, then, at the results of his survey: “The median answer of (mostly) financial professionals was that AI would NEVER exceed humans at all tasks,” he wrote in a post.

In total, 59% of those surveyed said AI would never overtake humans. Benartzi compared that to survey results published by AI Impacts of 3,000 AI researchers who are, presumably, both closer to the technology and its capabilities and more bullish on its potential. Of that group, only 15% are skeptical that AI will take over human tasks in the next 23 years; the vast majority believe it is likely.

The discrepancy between financial professionals and AI researchers shows “a big gap,” Benartzi says of his nonscientific poll. “I don’t have a crystal ball. But I think it’s plausible that in financial services, we don’t realize how fast this could happen. … We could be surprised, big time.”

Benartzi says using AI in the retirement industry is just another case of scaling technology to improve services for participants, as has been going on for decades with qualified plans. He also believes it will take great care and effort to get it right.

“AI has its own behavioral biases,” he says. “We have to learn how to use it in a way that is responsible.”

There is a misconception, however, that AI will run on its own and therefore not give the right advice or even give harmful advice, Benartzi says. Instead, he believes it can be guided by humans and act similarly to human advisers who have software and tools at their fingertips—only much faster and on a larger scale. In addition, AI will not be limited to areas of specialty when it comes to financial needs. It will be able to cover retirement planning, budgeting, investing, insurance, credit card debt, mortgages or whatever the person is seeking.

“The median 401(k) account, I believe, is around $16,000,” Benartzi says. “That’s not where you are going to add value. You are going to add value with debt management, with insurance choices, with health care insurance and all those important decisions.”

Benartzi agrees it will be challenging to make AI work in the ways he envisions, but says now is the time to get involved.

“Every challenge is also an opportunity,” he wrote in his post, “but only for those realizing it early on.”