Back to General Resources

House Financial Services Task Force Hearing on Artificial Intelligence

House Financial Services Task Force on Artificial Intelligence

“Robots on Wall Street: The Impact of AI on Capital Markets and Jobs in the Financial Services Industry”

Friday, December 6, 2019

Key Topics & Takeaways

  • Effect of AI on the Workforce: Fender noted that adaptation of the workforce is key, saying deep subject matter expertise combined with an ability to cut across disciplines will be essential as the industry changes. She said more research needs to be done to understand the emerging workforce trends in this space.
  • AI for Fraud Detection and Regulatory Compliance: Asked how AI is being used to detect fraud and unusual market behavior, Rejsjö said that Nasdaq relies on algorithmic coding to pick up on unusual patterns in the market. Wegner said it will be vital for regulators to have their own AI, saying this is an area where public-private partnerships are needed. She continued that as systems become more complex, regulators need to keep pace.
  • Bias: McIlwain said that as technological advances are made some groups of people are left behind and disadvantaged. He said that although technology is unpredictable, the exclusions that result are predictable. He continued that in looking at who is prepared to be part of the evolving technology sector, it is tremendously unrepresentative of all citizens of country, saying it is important to build a more inclusive workforce in these sectors going forward.

Witness

  • Charlton McIlwain, Vice Provost for Faculty Engagement and Development and
    Professor of Media, Culture, and Communication at NYU
  • Marcos Lopez de Prado, Professor of Practice, Engineering School, Cornell University and Chief Investment Officer, True Positive Technologies
  • Rebecca Fender, CFA, Senior Director, Future of Finance, Chartered Financial Analyst
    Institute
  • Kirsten Wegner, Chief Executive Officer, Modern Markets Initiative
  • Martina Rejsjö, Head of Nasdaq Market Surveillance, Nasdaq Stock Market

Opening Statements

Chairman Bill Foster (D-Ill.)

In his opening statement, Foster noted there are a number of ways artificial intelligence (AI) is being deployed in the capital markets, from automated trading to portfolio allocation to investment management decisions. He said it is also important to consider how the use of this technology is changing the nature of work in financial services, rendering some jobs obsolete and changing the skill set necessary to excel in others. Foster highlighted that computer-managed funds including index funds, ETFs and quant funds make up about 35 percent of the approximately $31 trillion American public equities market, while human-managed funds including hedge funds and other mutual funds make up 24 percent of the market. He said that while the rise of the “so-called computerization” of the stock market has benefits including lower costs to execute trades and more liquidity in the market, it also raises questions about algorithmic trading creating market volatility, such as the 2010 “flash crash.”

Ranking Member Barry Loudermilk (R-Ga.)

In his opening statement, Loudermilk said that in recent years there have been many technological developments, including the adoption of AI and automation, that have redefined and reshaped trading and investing. He noted that investors can trade securities from almost anywhere in the world using modern technology and electronic trading has immensely benefited the markets. He said electronic trading makes markets more efficient by allowing faster searches for prices, better processing of large data sets, more transparent price information, and lower overhead and transaction costs. Loudermilk highlighted challenges as well, including disruption to the job market. However, he said that although floor traders are being displaced, there is also more opportunity in fields such as code writing, cloud management, telecommunications, fiber optics and data analysis. He said it is important to keep in mind that not all types of electronic trading are the same, as there are differences between automated trading, algorithmic trading and high frequency trading.

Testimony

Dr. Charlton McIlwain, Vice Provost for Faculty Engagement and Development and
Professor of Media, Culture, and Communication at NYU

In his testimony, McIlwain focused on the implications of automation on the workforce and mitigating algorithmic discrimination and bias. He said there is ample reason to be concerned about the future of automation in financial services, noting that minorities already underrepresented in the financial services workforce will be most at risk in terms of automation, noting that the adjacent tech sector has similar underrepresentation issues. He said that to mitigate the likelihood that automation and innovation will negatively and disproportionately affect those already underrepresented, it is important for the industry to plan long-term. McIlwain said a way to combat algorithmic bias is to develop best practices for constructing and deploying algorithmic systems and provide more oversight to assess the outcomes they produce.

Dr. Marcos Lopez de Prado, Professor of Practice, Engineering School, Cornell University
and Chief Investment Officer, True Positive Technologies

In his testimony, Lopez de Prado said that as a result of recent advances in pattern recognition, supercomputing and big data, machine learning algorithms can perform tasks that until recently only expert humans could. He said that the key advantage of algorithmic funds is that their decisions are objective, reproducible and can be improved over time. Lopez de Prado explained that financial AI creates a number of challenges for the six million people employed in financial services and insurance, adding that many job losses will not be because they will be replaced by machines, but because they have not been trained to work alongside algorithms. He said that while machine learning algorithms can incorporate human biases, there is a better chance of detecting the bias in algorithms and measure it with greater accuracy than with humans because algorithms can be calibrated to perform as intended.

Rebecca Fender, CFA, Senior Director, Future of Finance, Chartered Financial Analyst
Institute

In her testimony, Fender described the impact of technology on jobs in financial services as a pyramid: at the foundation are basic applications that everyone must become more comfortable using, in the middle are specialist applications where technology will enhance work, and at the top are hyper-specialized roles that will be less common but very valuable, like those at quant firms and AI labs. She said AI unlocks the potential of unstructured data, can identify patterns in information more efficiently than humans, and can amplify an investment team’s performance, though it cannot replicate its creativity.

Kirsten Wegner, Chief Executive Officer, Modern Markets Initiative

In her testimony, Wegner said that in recent years, automated trading has led to much of the replacement of the exchange floor, noting that technology has reduced the cost of trading for the average investor by half. She said global competition to adopt the latest AI will make it more efficient in terms of speed, processing time, depth of data and cost saving. She said to expect a proliferation of “reg tech” as AI becomes increasingly valuable for firms and regulators to police the markets more efficiently. Wegner said that AI and automation can and should be a tool rather than a replacement for humans, noting that though some jobs will disappear, others will grow including in technology and fiber optics.

Martina Rejsjö, Head of Nasdaq Market Surveillance, Nasdaq Stock Market

In her testimony, Rejsjö said Nasdaq believes AI can be used to target fraudsters and market manipulators, noting that Nasdaq’s surveillance program has been using algorithmic coding to detect unusual market behavior. She said the current accessibility of the markets, the increase in players with the ability to deploy their own technology and the exponential increase in data quantities act as the perfect ecosystem for market manipulators to “hide amongst the noise,” adding that this increased complexity introduces new challenges for monitoring. Rejsjö said that while it can be difficult to capture new behavior and remain proactive rather than reactive, Nasdaq is continuously training its AI to produce more and more accurate output, and that by incorporating AI, they are sharpening their detection capabilities to safeguard the integrity of the financial markets and benefit investors.

Question & Answer

Effect of AI on the Workforce

Loudermilk noted that the adoption of AI and electronic trading can disrupt the job market and displace floor traders, though technology also creates demand for workers in other fields, asking the witnesses what job fields are growing. Fender responded that adaptation of the workforce is key, saying deep subject matter expertise combined with an ability to cut across disciplines will be essential. She said more research needs to be done to understand the emerging trends in this space.

AI for Fraud Detection and Regulatory Compliance

Loudermilk asked how AI is being used to detect fraud and unusual market behavior. Rejsjö said that Nasdaq relies on algorithmic coding to pick up on unusual patterns in the market.

Rep. Sylvia Garcia (D-Texas) asked about the uses of AI for regulatory compliance. Wegner said it will be vital for regulators to have their own AI, saying this is an area where public-private partnerships are needed. She continued that as systems become more complex, regulators need to keep pace. Garcia also asked how AI can help with anti-money laundering (AML) and suspicious activity report (SARs) filing. Wegner said that it is vital to work on this, noting that regulators and the private sector must work together to gather information on best practices.

Bias

Rep. Emanuel Cleaver (D-Mo.) asked how inclusive the technology is and how to ensure all people benefit from it. McIlwain responded that as technological advances are made some groups of people are left behind and disadvantaged. He said that although technology is unpredictable, the exclusions that result are predictable. He continued that in looking at who is prepared to be part of the evolving technology sector, it is tremendously unrepresentative of all citizens of country, saying it is important to build a more inclusive workforce in these sectors going forward.

Protecting Source Code

Loudermilk said it is important for the Securities and Exchange Commission (SEC) to only obtain proprietary trading algorithms when absolutely necessary and only with a subpoena, asking the panel to elaborate on why protecting source code is so important. Wegner replied that the “real lifeblood” of automated trading is the source code, saying it is the valuable intellectual property that firms are competing against each other with both domestically and globally.

Data Sets

Foster asked what range of data sets are being used by AI and machine learning applications. Lopez de Prado said that there are many combinations of data sets that are being used including credit card transactions, location data, satellite imagery, transcriptions from earnings calls, engineering process data and others. He noted that 80 percent of all data ever created has been generated in just the last three or four years. He said data can be used to understand the psychology of people, and its use is only going to increase because it is becoming cheaper to store and processing power is increasing.

Monopolies

Foster expressed concerns about monopolies developing and smaller players being driven out of the market. Lopez de Prado responded that there are a number of academics who believe that a consolidation is not necessarily negative as the few survivors will operate more like utilities in the future. He noted, however, that a smaller number of operators could cause a domino effect if one of them fails to provide liquidity.

For more information on this hearing, please click here.

For an archive of past SIFMA hearing coverage, please click here.