The lead article in this issue of Inside Information talks about all the marvelous new capabilities of AI in our fintech space, and it’s fair to wonder how far that will go. Hey, Alexa, play some of my favorite Taylor Swift songs, and then please look at my income and last year’s tax statement, and check out my IRA balance. Do you think I should do a Roth conversion this year, and if so, how much should I convert to fill up our 24% tax bracket based on your best estimate of our taxable income this year?
Yes, I do think that’s a bit far-fetched; you have to know how to ask very sophisticated questions in order to get good planning advice out of an entity that still gets confused about whether a line drawing represents a rabbit or a squirrel—and most consumers won’t possess nearly that level of knowledge. But still… Alexa might eventually be trained as a financial planner, in which case the little machine listening to your every word from the living room end table would be asking the questions and proposing strategies.
How long before AI starts to compete with you and the advisor down the street?
I won’t say never, but I do think that’s probably the right answer. And the reason is not expertise or even relationships. The reason is trust.
To see why I believe this, let me take you on a brief trip to the future. It won’t be far; just to around 2032 or so, the year when we experience the first real AI scandal. The headlines will tell us how some large sales entity—probably a Wall Street wirehouse, but I can’t be sure—will build a very plausible, charming financial planning agent whose brain is packed with the contents of every CFP textbook, all million pages of the tax code, a superior planning calculator and a bedside manner that oozes charisma and calm reassurance. The voice will sound like God, only friendlier.
The only tiny fly in this particular ointment—which triggers the scandal—is that this wonderful AI agent that sounds like God will have been programmed to sell products that are not, shall we say, in the best interests of the person who is taking all this wonderful advice. And the only reason I think this will be a headline in 2032 instead of 2028 is that it will take a few years for the creator of this agent to get a bit greedy with all the sales success they’re having, and start to push out ever-more-noxious products whose malodorous aroma finally catches the attention of regulators, press and litigation attorneys.
The ensuing scandal will destroy public trust in all the AI agents the way Bernie Madoff made consumers suspicious of even fiduciary advisors. At that point, flesh and blood human advisors who don’t sound like an immortal deity will resume their eternal role as protectors of their clients from all the many sales agendas that all of us are routinely exposed to. People will realize that AI can only be trusted to the extent that you can trust its developers, and they will look to their human advisor to provide the trusted verification that the advice they’re receiving is in their best interest.
As they always have. As they always will—at least until 2040 or so, when somebody finally finds a way to give AI a workable conscience.