Finding the Goldilocks Rules for Artificial Intelligence

The Supremes said you can’t hurry love, and it turns out you can’t hurry regulations for artificial intelligence either. At least according to Federal Reserve Board Governor Michelle Bowman.

In a recent speech on “Artificial Intelligence and the Financial System,” Bowman offered a slower approach to oversight of AI amid growing calls for governments around the world to come up with ways to regulate the technology.

“We cannot adapt a one-size-fits-all approach as we consider the future role of AI in the financial system,” Bowman said. “Over-regulation of AI can itself present risks by preventing the realization of benefits of improved efficiency, lower operational costs and better fraud prevention and customer service.”

Bowman cautioned that because we live in a world where technological innovation moves at a faster pace, any regulatory regime for artificial intelligence “should be nimble” and financial regulators “must have an openness to the adoption of AI.”

But before any of that can happen, regulators and oversight boards need to learn more about AI, according to Bowman.

“We must understand AI before we consider whether or not to change our regulatory approach,” she said. But that implies the Fed and other financial oversight agencies have a regulatory approach right now.

Most of the rhetoric coming out of Washington, D.C. and other national capitols regarding AI is little more than proactive-sounding verbiage indicating a recognition that artificial intelligence is something governments should be aware of, and likely requires cooperation and collaboration in creating a regulatory framework.

Hence Bowman’s go-slow approach to federal oversight of AI in the financial system.

“We need not rush to regulate,” she said. “It is important that we continue to monitor developments in AI and their real-world effects.”

That sounds like a sensible approach. Get the facts, understand the issues and fashion regulations that protect consumers and the financial system while not stifling innovation. It almost sounds like something you would find in the mission statement of any of a number of regulatory bodies. Finding the balance between too much regulation and not enough – like porridge that is “just right” – is the eternal challenge of governments.

Too little oversight, and consumers, business owners and anybody else interested or impacted by the growth of AI are left to the mercies of the innovators and disruptors, with little recourse if things go terribly wrong. Too much regulation and AI and its potential for game-changing effects on banking, finance and society in general will wither on the vine.

But governments can’t take too long to figure out what they are going to do. Scientific discovery and technological change wait for no man – or woman. And if regulators sit on their hands trying to find that just-right solution, it could end up being too late.

Perhaps the most prescient line in Bowman’s speech both supports and undercuts her advice to take things slow on the regulatory side: “We will almost certainly be surprised by how (AI) is ultimately used.”

She’s right, of course. We may just not like the surprise.