“It would not be an exaggeration to say today that Wall Street is – quite literally – run by computers.”

Congressman Bill Foster, Chairman of the Committee on Financial Services Task Force on Artificial Intelligence made this definitive statement at a recent hearing exploring how AI is impacting the financial services industry.

Wall Street is Run by Computers

Hearing participants included:

While the hearing focused largely on how AI affects investment decisions and implications for the financial services workforce, reg tech was also a topic of discussion. 

In other words, how can automation increase regulatory oversight, assist with risk management and help firms with compliance?

The Committee, and the hearing participants, all noted that AI will help make the regulatory processes more effective and efficient.

A Proliferation of Reg Tech

“We can expect to see a proliferation of reg tech as AI becomes increasingly valuable to police the markets more efficiently,” said Wegner. “AI functionality in reg tech includes monitoring, reporting and compliance, and processing of regulatory filings, loan origination processing, detection and reporting of illegal and irregular trading and detection of cyber risks.”

In its Memorandum related to the hearing, the Committee wrote that “AI has the potential to make compliance activities, like anti money laundering (AML) compliance, more accurate and efficient.”

Millions of SARs Filed Annually 

Consider that when it comes to AML compliance, financial firms file millions of suspicious activity reports (SARs) with the U.S. Treasury’s Financial Crimes Enforcement Network (FinCEN)  each year. 

It’s then dependent on FinCEN regulators to determine which of those reports are actionable.

That’s an incredible amount of manual work that AI can assist with, by using algorithms to sort through the data filed and help prioritize the reports.

Tackling Petabytes of Data

“We have to tackle tremendous amounts of data – petabytes of data – and identify this needle in the haystack. I think a practical solution is for regulators to work together with data scientists and with the entire community and crowd source these problems,” said Lopez de Prado. “It would be difficult for the agencies to develop the kind of techniques that the wrong doers are developing for bad purposes and number two, the amount of data to parse through.”

An Effective Cop on the Beat

Wegner agreed, saying “As bad actors become more sophisticated globally, it’s absolutely vital that financial regulators have the funding resources so they too have the technological capacity and access to AI and automated technologies to be a strong and effective cop on the beat.”

In fact, Nasdaq is already using AI to help with its market surveillance, looking for insider trading, fraud and manipulation as well as erroneous transactions.  The surveillance program uses 40 different algorithms that have 35,000 parameters to detect issues in real time.

AI will Play a Key Role in Detecting Manipulative Behaviors

“By incorporating AI into our monitoring systems, we are sharpening our detection capabilities and broadening our view of market activity to safeguard the integrity of our country’s markets….Moving forward, AI will play a key role in detecting manipulative behaviors that would otherwise undermine our markets,” said Rejsjo.

Rejsjo was also clear that the AI work being down by Nasdaq included human input via “active learning.” 

The importance of keeping a human in the loop was echoed by others on the panel.

AI + HI: It’s Not a Race Between Humans and Machines

“We believe in the power of the ‘AI + HI’ model – that is, most tasks are and will remain best handled using both AI and human intelligence, and the collective power of the two is superior to either element on its own,” said Fender. “It is not a race between humans and machines. The competition ultimately is among ‘AI + HI’ teams, and the stronger teams that effectively harness and combine both elements will outlast the weaker ones.”

Another benefit of AI is its ability to help mitigate inherent human bias.

Mitigating Human Bias

“Yes, machine learning can incorporate human biases. The good news is that we have a better chance of detecting the presence of biases in algorithms and measure that bias with greater accuracy than on humans.  The reason is that we can subject algorithms to a batch of randomized control experiments and recalibrate those algorithms to perform as intended,” said Lopez de Prado. 

He went on to share the critical part policymakers will play as AI continues to be integrated into regulatory processes: “Congress and regulators can play a fundamental role in helping reap the benefits of this technology while mitigating its risks.”

Learn more about AI-powered Radiance and its capabilities to support KYC, AML and other regulatory compliance issues for the financial services sector.