The explosive growth of artificial intelligence is shaking up every industry, including financial services. Business investments in AI are “forecast to approach $200 billion globally by 2025.” By 2030, “AI could contribute up to $15.7 trillion to the global economy,” with a boost in productivity potentially accounting for $6.6 trillion. In the financial sector, spending on AI is predicted to “more than double to $97 billion in 2027.”

The hype around AI is inescapable. But as with any hot new technology, the hype—fueled by misconceptions, memes and half-truths—can often get in the way of the facts. I’m excited and optimistic about how we can use AI to improve investment management, but I also think it’s important to understand what it can and can’t do. Let’s take a closer look at three common myths and explore the practical applications and benefits of AI in trading.

Myth: All AI Is Created Equal

Since ChatGPT launched in November 2022, the buzz around generative AI has been deafening. Generative AI is already making major contributions in diverse business applications, from customer service to content creation. Generative AI is also notoriously bad at math, which leads many people to lump all AI together and think that AI serves no purpose in investment management.

But generative AI and predictive AI are very different entities with distinct strengths and weaknesses. Generative AI summarizes and synthesizes available data that it has been trained on, producing human-like responses to queries and creating new content, including text, images or music. Predictive AI, on the other hand, analyzes large existing datasets to identify patterns and make projections or recommendations about future events or trends.

Myth: AI Is All Or Nothing

If you ask generative AI to create a portfolio or make investment decisions, you will quickly see there is no danger of it replacing a human investment manager. It collects information from online sources that use non-standard data and for which there are no hard-coded boundaries, so you can't be sure its answers are accurate or complete.

For example, if you direct a generative AI system to gather all of the 13F and 13D quarterly filings for investors who have expanded their investment in eight specific stocks by more than 5% over the last 15 years, you wouldn't be able to trust the results. Without manually verifying the data, how could you be certain the system captured all the filings? Would it know that Form 13F was different 15 years ago than Form 13F is today?

But predictive AI excels at analyzing vast volumes of historical data and making educated projections about future outcomes. It collects verified data at the source, compares it with other inputs and identifies specific patterns and insights. You can buy data from multiple sources, such as NASDAQ, Opera and ICE, and cross-check it for errors. My investment management firm exclusively uses predictive AI that works within clearly defined parameters. Because it's capable of billions of simultaneous examinations, its pattern-recognition abilities are vastly superior to those of a human.

Predictive AI is not a crystal ball, but if you understand how to use it effectively, it can be a powerful investment management tool. If you’re a fundamental investor, you wouldn’t take predictive AI's discrete suggestion to buy a breakout stock—but you might trust its forecast that a market is likely to fail and be more conservative in your allocations as a result. If you’re a quantitative investment manager, you might have predictive AI take over many of your manual tasks. Don’t look at AI with a black-and-white, all-or-nothing perspective. You can find ways to use it to your advantage.

Myth: Autonomous AI Is Completely Autonomous

When people hear about autonomous AI, they often jump to worst-case scenarios of machines making decisions without human oversight. But you don’t have to worry about Skynet or HAL 9000 becoming self-aware and running amok. Autonomous AI doesn't have complete autonomy. It is capable of making limited decisions within a framework you create. Driven by reinforcement learning, autonomous AI in trading adapts to changing market conditions and new data, while still adhering to the risk protocols and trading strategies you define.

For example, if you use 26 different criteria to make trade decisions, you could allow AI to reorder but not add or remove criteria. As AI observes billions of iterations of your strategy, it is constantly analyzing and validating how you weight these criteria, and it can move No. 15 to No. 1, but it can’t remove No. 23 altogether. There are always guardrails in place.

AI is not going away—and I strongly encourage you to see it as an opportunity and not a threat. Don’t let fear or unfamiliarity hold you back from using this transformative technology to your benefit. Instead of worrying that AI is coming to take your job or thinking you have to learn how to program to use it, start building a healthy, positive relationship with AI. Ask: How can I use this to my advantage? There is no one right way to use AI for investment management. It can do most of the heavy lifting for you, or it can just do some of the annoying tasks you don’t enjoy. Think of AI as a new partner with limitless potential, and stay open to what it can do for you.

LeackStat 2024