Some advisors have adopted artificial intelligence (AI) for administrative functions such as transcription, client management and data analysis, but many are reluctant to rely on the technology for portfolio management.
However, clients are showing an interest. According to research from the Ontario Securities Commission (OSC) and The Behavioural Insights Team (BIT), Canadians are more open to following advice from an AI tool.
It’s clear confidence in AI is growing, says Sasha Tregebov, BIT’s director in Toronto. That’s happening despite people having this “innate psychological, even non-conscious, preference for human judgment,” he says.
When they began their research, he says BIT and the OSC believed they would see an aversion to AI advice. The randomized controlled trial tested how participants followed advice about investing $20,000 in fictitious funds. Participants were told the recommendation came from either a human advisor, an AI tool or a human using an AI tool.
“What we found is that people most closely followed the advice of a human using AI,” Mr. Tregebov says. “AI was second place, and the human was third place.”
Mr. Tregebov says clients are ready for AI to manage their assets. The question is whether advisors are.
Jason Pereira, partner and senior financial planner with Woodgate Financial Inc. in Toronto, doesn’t believe they are. He says AI-driven research used to create customized indexes and portfolios based on specific parameters set by clients, which “give you a portfolio of XYZ stocks that would meet those objectives,” have been around for a long time. But he says the move to portfolio rebalancing is yet to come.
“A lot of the software exists that can actually do that, but I don’t think anyone’s willing to turn the keys over just yet,” he says.
AI has to become more sophisticated before it’s used in portfolio management, says Thane Stenner, senior portfolio manager and senior wealth advisor with Stenner Wealth Partners+ of CG Wealth Management in Vancouver. He says a lot of the data online that’s synthesized by AI tools is consensus view – and “consensus view isn’t necessarily the right view.”
But that’s changing, and Mr. Stenner believes that with improved accuracy and refinement, AI will lend itself to “a lot of applications.”
“Our analysts and our group use AI, but you still have to eyeball it, assess it, and use some intuition to see what’s what,” he says. “Because at the end of the day, it’s junk in, junk out.”
Mr. Pereira believes the adoption of AI in portfolio rebalancing will come down to trust. He likens the current situation to a “driverless car problem.”
“By all accounts, the data show that driverless cars are safer than human drivers at this point,” he says. “But how many human beings are willing to turn over their lives to a driverless car right now? The reality is, we are far more willing to accept errors from humans even if the technology’s error rate is substantially lower.”
He says AI-driven systems currently exist that are more likely to rebalance portfolios at the exact right time – “maybe intraday, as opposed to the end of the day” – than humans are. “But the reality is, at this stage, not many people want to take that risk.”
He says legal liability is also a big concern. If the software firm making the AI-enabled tools used to make calculations was to take on the legal liability of providing financial advice and stand behind the results it generates, that would absolve advisors of liability and encourage them to use it.
Mr. Stenner says that as AI improves, advisors’ perspectives will change.
“It’s going to get better and better,” he says, adding that, for now, advisors still have to watch things very closely to ensure accuracy.
Mr. Tregebov agrees that advisors need to continue to be the gatekeepers to safeguard their clients.
“People want advice, they listen to advice, but they’re not really that great at discerning between good advice and bad advice,” he says. “That raises the stakes for the industry, for the regulator, for all of us.”