The possible upside from financial sector innovation may be compelling, but regulators must be wary of the risks, argues a new report from the C.D. Howe Institute.

The Toronto-based think tank issued a report that reviews the history of financial innovation and highlights lessons for policymakers today.

While innovations such as “open banking” and the use of artificial intelligence (AI) may benefit both consumers and the industry, the report called on regulators to be vigilant to the stability risks that may arise.

In particular, the report said innovation may produce easier credit and generate a credit supply boom.

“This is likely to be more problematic if the boom is in household mortgage credit,” said the paper’s author, David Longworth, adjunct professor in the department of economics at Queen’s University, and former deputy governor at the Bank of Canada.

He also warned that the risk is higher if the increase in credit comes from the so-called shadow banking sector.

The report said “regulators need to take seriously the fact that they have little or no data on non-bank financial intermediaries (NBFIs), or shadow banks.”

In response, it recommends that regulators collect comprehensive data on lending and short-term financing.

“Special attention should be paid to lending from Big Tech companies — such as Google and Amazon — especially to small and medium-sized businesses, as well as online, peer-to-peer lending for both households and businesses,” it said.

And, it recommended that financial sector stress-testing also include the risks posed by shadow banks that are not prudentially regulated.

“Such stress tests should take into account, for instance, substantial new financing in the form of uninsured mortgages from these institutions,” it said.

The report also said regulators should ensure adequate oversight of AI-based lending decisions by “requiring the use of Explainable AI (XAI), which employs techniques that allows human oversight of the AI’s decision-making process.”

It also noted that innovation could increase the possibility of runs on short-term funding. For instance, apps “could increase the likelihood of bank runs in response to errors banks make in posting their interest rates or to errors the apps make in interpreting the information.”

The recommendations, which aim to preserve financial stability, shouldn’t have any negative effect on most financial innovations, the paper said.