The latest in enforcement: robo-experts

Forget robo-advisors. The latest in tech is the robo-expert. At least it could be, if the Monetary Authority of Singapore (MAS) is successful in its latest endeavour.

Gillian Tan, executive director of enforcement at the MAS, spoke this week at the OSC’s Dialogue 2017 at a session on the latest in enforcement.

Read: Why fintech has an incentive to be regulated: BoC paper

She said the MAS plans to use artificial intelligence “to get a machine to replicate an expert’s thinking.” Essentially, robots will be trained to analyze data, including past expert reports. These robo-experts could potentially be used as court witnesses.

Overall, the session’s message was that, with data analytics, enforcement is becoming proactive and streamlined.

With limited resources, enforcement traditionally focused on only high-priority cases, said Tan, and data was analyzed manually. Now, enforcement does more with less.

“Analytics and technology — if harnessed correctly — bring us huge efficiency gains,” she said.

For example, behavioural data models are now applied to trading data to identify collusive price manipulation and circular trading, said Tan, adding that the MAS was motivated to use tech because of syndicated market manipulation. What used to take months, now takes hours, she said.

Stephanie Avakian, co-director of enforcement at the SEC, said the commission uses tech tools to analyze data and subsequently identify suspicious trading, trading patterns and trader connections.

While the new face of enforcement is tech-based, said Jeff Kehoe, the OSC’s enforcement director, tech tools must be reliable, predictable and fair. “We can’t be tempted to overuse or rely exclusively on these tools,” he said. “There has to be judgment. […] Humans are still in charge.”

Those humans might not be the usual professionals, however. Kehoe said more data scientists are needed — and fewer lawyers and accountants.

Tech as deterrent?

“Are tech tools effective deterrents?” asked an audience member.

We don’t know, said Avakian, because there’s no way to measure. However, a positive development is that registrants are getting onboard with their own data analysis, she said.

If we can take some of the work upstream, there will be fewer problems downstream for regulators to resolve, said Tan. Further, making enforcement faster with data analytics is a deterrent in itself.

Matt Cardillo, senior director of market regulation technology at FINRA, said data triangulation allows risk profiling of firms. Past trends can predict the future, he said.

Also read:

OSC’s priorities provide clues to enforcement

Don’t forget the basics as fintech adoption booms

How big data can make insurance better