This is part 2 of a roundtable Benzinga conducted on Wall Street’s adoption of artificial intelligence. Click here for part 1.
Benzinga asked a panel of people specifically working at the intersection of AI and Wall Street to give us their thoughts on how the trend will play out. The panelists are:
- Alex Lu, co-founder & CEO of Kavout, an AI firm that provides quantamental analytics
- Chris Natividad, CIO of Equbot, provider of the AI Powered Equity ETF AIEQ
- David Aferiat, co-founder & managing director of Trade Ideas, an AI-powered discovery engine
- Dr. Hossein Kazemi, Ph.D. and Senior Advisor to the CAIA Association
- Lane Mendelsohn, vice president of VantagePoint, an AI market forecasting software
- Nikhil Dhingra, co-founder & CEO of Tradagon, a fintech company that uses AI in its signals.
Some answers have been edited for clarity.
4) How far ahead or behind is Wall Street in implementing AI compared to other major industries?
Alex Lu, Kavout: I wouldn’t label Wall Street as a laggard in adopting AI. Internet industry definitely stays at the top regarding the use of AI. But Wall Street definitely takes one of the top three spots and leads many other industries when it comes to implementing AI in business.
Chris Natividad, Equbot: The finance industry is definitely behind compared to other industries. Technology adoption practices and regulatory requirements throughout the finance industry are just a couple of operating aspects causing friction in implementing AI.
Technology is the obvious industry leader and is signaling a seismic shift with significant investment. IBM IBM, Google, Microsoft, Intel Corporation INTC, NVIDIA Corporation NVDAO and Amazon.com,AMZN are all bolstering their AI research and involvement. We believe as more instances of AI success surface the overall adoption rate outside of the technology industry will increase.
5) What's the biggest challenge that AI will cause for regulators?
6) Will the adoption of AI lead to a reduction in volatility?
Dr. Hossein Kazemi, CAIA: My guess is that the asset management industry is somewhat behind the transportation and information technology industries.
Lane Mendelsohn, VantagePoint: I think Wall Street is behind other major industries. I think the healthcare industry has really adopted artificial intelligence and has made some great strides. I know that a lot of telephone companies and customer service-based companies are using AI to help them reduce hold times and get answers to questions that people have much more quickly. A lot of other industries have had an easier time figuring out what the problem is and how AI can solve that problem, and then the implementation is much easier.
Nikhil Dhingra, Tradagon: I would say it is not different than the Financial sector adopting any other new technology...it seems it moves somewhat slower due to more red tape, regulation, and just longer sales cycles in general.
Alex Lu, Kavout: If you take a deep dive into AI, you will realize that its foundation is machine learning, a domain resting upon mathematics and statistics, just like the well-known quant or statistical models in Wall Street.
The confusing part originates from its presentation or positioning as “intelligence” versus tools, or analytics. Once something is labeled as “intelligence”, we intend to think of it as another “being” or something with independent consciousness. This drastically increases the chance that regulators will have the obligation or curiosity to ask for more details and explanation of where that intelligence comes from and how we could quantify and control the financial risk associated with AI.
Unfortunately, a machine learning model with a fair amount of complexity works like a black box and can’t be 100 percent explained even by the very engineers who build it. This lack of explanatory nature of AI needs to be addressed by practitioners, while regulators need to keep an open mind to be adaptive, invest time to understand the new paradigm and changes, and leave AI more space to grow.
Chris Natividad, Equbot: AI can greatly benefit regulators, so their willingness and ability to adapt will be important. Automation through machine learning could potentially make these agencies more efficient and allow them to focus on critical business areas. Innovation at regulatory agencies may help them better prepare or anticipate potential industry developments.
David Arferiat, Trade Ideas: I don't see the challenge. If anything, AI will allow, in one aspect, for a smoother discussion of Facts & Circumstances behind a trade that a government regulator would want to know about. The AI details around key trading decisions would be enough, if properly explained, transparency to understand why a trade was made. Access to the machine learning rule sets would be what's required.
Dr. Hossein Kazemi, CAIA: There are many, but one particular challenge of note to our work at the CAIA Association is around competency. For example, would regulation require a robo-advisor’s “robot” to pass the Series 7, 8, etc. How about the CPA exam if the “robot” is going to do taxes?
Another example is that AI’s results may lead to discriminatory results, but legal recourse might be difficult. There is a study that shows that when AI is used to approve mortgage loans the results tend to discriminate against minorities. AI might be good in finding association, but not good in finding causation. That is, we could have meaningless learning in machine learning.
Lane Mendelsohn, VantagePoint: I don’t really know how artificial intelligence will be a challenge for regulators. I think that’s just something that we’ll have to wait and see how AI continues to develop and change the world specifically related to trading and investing before we’ll know how that will affect or be a challenge for regulators.
Nikhil Dhingra, Tradagon: Probably security and creating hard and fast rules around something that is still very novel and, some would argue, requires a Ph.D. to actually understand.
Alex Lu, Kavout: I have not seen any statistically significant analysis stating that the use of computers in trading increases or reduces the market volatility compared to pre-computer age. If we view AI as just another technology, it does not have much to do with market volatility.
But if major institutions all use the homogeneous AI system, very likely we will have a more fragile financial market. Just like a species in nature with little diversity in its gene pool, this species has a much higher probability to die out with an unexpected virus. So is the financial market. But again, this is a thought experiment with a lot of my assumptions. The reality will tell the truth in the next 5-10 years as more and more institutions adopt AI. As long as everyone uses different algorithms and parameters, we will not see a lasting impact of AI on market volatility.
Chris Natividad, Equbot: We believe initially there will be greater volatility as a technology gap will differentiate industry players. The evolution of AI will likely further reduce the window of time to capture mispriced opportunities in the market as the technology will drive more efficient security selection and trading.
We see the massive amounts of data AI can process in an unbiased manner and the ability of the technology to derive new investment efficiency metrics. We envision this market volatility eventually moving toward a more normalized state as the industry evolves into a more AI friendly paradigm. However, that "normalization" will last only until a new technology arrives.
David Aferiat, Trade Ideas: AI is just modeling the market's behavior so there's no reason that it would add something that's significantly different than current volatility ranges. Additionally, AI-generated analytics are a different tool for trading but not nearly as impactful on liquidity as auto trading/HFT. If anything, AI stands to reduce volatility if it brings in more people to trade in the market—more participation in the markets generally lowers volatility.
Dr. Hossein Kazemi, CAIA: In the short run, I think it will reduce volatility as some of the traders who are in crowded trades may move to less crowded ones due to the use of AI. That is, correlations among trading strategies could decline in the short-run. However, if AI becomes widespread and all AI programs discover the same trading opportunities, then we could have many more flash crashes. These are all questions that CAIA and the CADA Institute will continue to study.
Lane Mendelsohn, VantagePoint: I don’t think that the adoption of AI will lead to a reduction of market volatility. I think the markets are always going to be volatile, and we need volatility. Volatility is actually a good thing for the markets. It creates tremendous trading opportunities for people to really build wealth.
The key is the AI, and this is something that we’ve successfully been able to do for decades now, is the AI can help traders navigate the market volatility. And that’s really what we’re looking for. If there was no volatility, then there’s not really much in the way of trading opportunities. If there’s volatility, that creates trading opportunities, but the only people that will be able to seize those opportunities and profit from them are those who have a way of navigating that market volatility and putting the odds in their favor. That’s how VantagePoint has helped our customers utilize artificial intelligence to do just that.
Nikhil Dhingra, Tradagon: I think initially definitely volatility simply due to rogue algos misfiring. And then, who knows...one hopes that it does not make volatility something of the past.
For part 1 of this roundtable, click here.
© 2023 Benzinga.com. Benzinga does not provide investment advice. All rights reserved.
Your update on what’s going on in the Fintech space. Keep up-to-date with news, valuations, mergers, funding, and events. Sign up today!