At Generative AI World, we heard from Dr. John J Sviokla, co-founder of GAI Insights and former business leader at PwC; here are three of his main...
Lisa Huang spoke at Generative AI World about the future of finance. She has a lot of experience under her belt, for example, starting out at Goldman Sachs, and her involvement in the Betterment robo-advisor project, and now her role at Fidelity as Head of AI Investment.
Here are three of the big themes from Huang's talk.
- She separates the future into four categories – the present, with what she suggests is a “sprinkle” of AI, the near future with automated AI, the not-too-distant future with DeFi, and the more distant future with a likely focus on quantum computing.
- LLMs, she says, can give good investing advice, and even explain why they're giving it. That comes from her personal research with these models.
- Graphs are important technology for solving unique problems in finance, some of which are related to crowdsourcing and mass investor perspective.
“I think a lot about the future of assets and money – and I think DeFi has the capability to fractionalize any asset, and with fractionalization, that makes everything very cash-like. It provides a lot of liquidity.”
“There is a set of challenges when you try to apply AI to investing – markets are not stationary … it's not as sticky as human behavior – when you order from Amazon, you tend order the same things.”
“When we first started this work, we were in discussions (about how to tell when a security is crowded) – somebody in the room said: ‘there is no definition of crowdedness – it's an emotion – how do you differentiate when you are having an emotion? (But) when you put in a graph – it's actually really easy.”
More from Lisa Huang:
Investing, she says, has a small data set problem.
“You need a lot of confidence (for some of this analysis)” she says, noting that some of these goals are “not insurmountable” as the technology matures.
She identified three big questions for investors, fundamentally: 1. what to buy 2. how much of it and 3. When. AI, she suggested, helps with all three.
In brainstorming solutions, Huang talked about “asset allocation through collective intelligence” and later warned of a “behavior gap” which can lead to a portfolio that underperforms by up to 1.11%.
The future, Huang predicted, is LLMs that can give you financial advice at scale. For particular value adds, she went over potential for optimal tax strategy, and the use of personal indices.
The traditional rule is that the S&P is the benchmark, she said, but it doesn't have to be.
Huang talked about wealth management “infused with AI design” and described how, in the future, the technology will be the investor, planner, therapist, educator and coach.
Get the full video of Huang’s speech at the conference with a Basic Membership ($89/year) to GAI Insights.