Smarter Portfolios: The Risks and Benefits of AI-Powered Retail Investing

There are several types of AI tools that people use for investing. Many trading apps have AI-powered analytical tools built into their dashboards to crunch data for users; for example, Public.com uses an AI assistant to look up stock data.

However, when people talk about AI’s growing influence on retail investing, they’re usually referring to more than just advanced stock lookup tools. Most of the time, they mean one of these types of AIs:

• Large language models (LLMs): Chatbots like ChatGPT, Gemini, and Claude are all large language models, also known as generative AI models. Like Stuart, many people have begun asking these AIs for tips about investing and personal finance, essentially treating them like free financial coaches.
• Robo-advisors: Many investing platforms (like Betterment and Charles Schwab) also offer “robo-advisor” software, which builds and manages investment portfolios for their users based on their risk tolerance and goals.

LLMs and robo-advisors are both considered types of AIs, but they work in very different ways.

Robo-advisors use prebuilt algorithms to create investing strategies. Although these algorithms are sophisticated, they aren’t anything new; robo-advisors have been around in essentially their current form since the 2000s, although the underlying algorithms have grown more advanced since then.

LLMs, on the other hand, are newcomers to the finance world. They entered the popular consciousness with the release of ChatGPT in 2022, and have taken the world by storm since then.

They’re essentially text prediction tools, although they’re very advanced ones that are, at this point, capable of a certain degree of genuine analysis. They are, for instance, capable of evaluating the overall sentiment around a stock and judging whether most people think it’s likely to rise and fall in price.

Both robo-advisors and LLMs are widely used, but because LLMs are so much newer, most recent stories about AI and investing have focused on them.

How many people are using AI for investing?

According to a recent Investing.com survey, 62% of retail investors actively use AI tools to guide their decisions.

There appears to be an age divide in the popularity of these tools. The eToro Retail Investor Beat study found that roughly 69% of Gen Z investors are using AI, compared to just 35% of Baby Boomers.

When it comes to the use of generative AI, specifically, the enthusiasm is similarly concentrated in younger generations. A survey by CreditKarma found that 66% of Americans are using these kinds of platforms to seek financial advice, but the percentages for Gen Z and millennials are 82% each.

Most people aren’t outsourcing their investing decisions to bots like GPT yet the way Alexander Stuart did, but a significant number—around 19%, or one in five people—have done so, according to eToro. More than double that (39%) are at least considering the idea.

Stuart’s story proves that sometimes, using AI to make investment decisions can turn out well, and many financial professionals acknowledge that AI can be a valuable educational tool.

As early as 2023, when the technology was still in its relative infancy, a popular financial blog, the White Coat Investor, tested GPT and found that while its advice wasn’t perfect, it also “didn’t suck,” and that the AI was surprisingly willing to admit the limitations of its own guidance.

In 2026, Jill Schlesinger of CBS News acknowledged that AI can be “great for general education” and that it’s good at answering questions like “how do I read my pay stub?” and “what’s the difference between a Roth and a traditional IRA?”

However, she added a caveat: “When you’re making a big decision, it’s kind of scary to rely on AI, which can hallucinate.”

Fake and obsolete data

“Hallucination” refers to LLMs’ infamous tendency to confidently state incorrect or out-of-date information.

Language models do this because they’ve been designed to generate satisfying-sounding answers for their users, which isn’t necessarily the same thing as generating correct answers.

Although AI companies like OpenAI and Anthropic have been trying to teach their models not to do this, the tendency has proven hard to eradicate.

Even Alexander Stuart, one of AI-driven investing’s success stories, found this out the hard way. The New York Times reported that one month into his investing journey, he “lost nearly $60 on an Nvidia trade”—a substantial portion of his relatively modest portfolio—because GPT mistakenly displayed a number that was several days old.

The Federation of Small Businesses, a UK-based organization, compared using AI to hiring an overeager intern who’s desperate to impress. “If you ask for a market summary, they will write a beautiful one. But if you ask for a quote or a statistic that doesn’t exist, they might just invent one to complete the task.”

Put another way, when used as research aids or educational tools to round out your financial literacy, AIs can be valuable, but when it comes to using them to pick specific stocks, look out.

A CFA’s perspective

We asked AI engineer and Chartered Financial Analyst Mehdi Zare about the wisdom (or lack thereof) of using chatbots to pick stocks.

He affirmed the existence of these pitfalls. Investors can easily get misled when they pitch investing ideas to an AI, he said. “The bot is built to be agreeable … and they walk off feeling validated by a ‘yes man.'”

He also expressed concerns about chatbots’ handling of time-sensitive data. “The base model’s knowledge is months stale, so if you ask for a price or a recent earnings number, it’ll hand you a confident wrong answer.”

Another problem is that you can’t sue chatbots or report them to a licensing board if they squander your life savings. Unlike many human financial advisors, they have no fiduciary duty—a term for a legal obligation to act in your best interests.

AI-related financial losses are common

Although not everyone who turns to AI for investing advice runs into these issues, many do. A 2025 survey from an AI firm called Pearl found that 19% of Americans (and 27% of Gen Z investors) have lost more than $100 in the stock market after following advice from LLMs.

Investopedia reported on the survey and quoted Jared Gagne, a CPFA at Claro Advisors.

“AI can give decent general advice that turns into very bad advice when applied to the wrong situation,” he said.

This isn’t just limited to inadvisable stock picks; he also cited the example of an AI advising a trader to cash in stock options, a kind of financial instrument, in a way that could trigger a huge and unexpected tax bill.

Defenders of chatbots like GPT might object—with some justification—that other AI tools come with problems as well.

At this point, robo-advisors are considered time-tested tools, and they aren’t particularly controversial. However, investors using them have been burned, too.

In 2022 and 2026, the Securities and Exchange Commission fined Charles Schwab and Ally Financial $187 million and $500,000, respectively, for allowing their robo-advisors to mislead clients.

These institutions’ investment algorithms automatically locked up to 30% of their users’ portfolios in cash so that they could lend that money out. Consequently, many investors lost out on gains when the stock market rose in value.123

What experts think about robo-advisors

Like LLMs, robo-advisors get mixed reviews from finance professionals and economists. They can save investors time by scanning financial data and automatically rebalancing their portfolios, and they can also curb emotional trading driven by fear and greed when markets get volatile.

However, it’s hard to know how they’ll perform during economic storms. One study from 2023 found they yielded 12.67% better returns than human investors during the pandemic-induced market crash. Conversely, a 2026 study published by the European Central Bank found that AI systems yield unpredictable results for individual portfolios during market instability.4 Ultimately, the jury is still out on how reliable robo-advisors are.

Of course, it’s more than possible for things to go sideways when retail investors take matters into their own hands, too, and the same goes for when they work with (human) financial advisors. The ultimate lesson might simply be that investing is inherently risky, and using AI doesn’t magically turn it into a sure thing.

Like many of the other experts we’ve quoted, Zare told us that he recommends using chatbots for educational purposes rather than treating them like stock-picking tools.

For example, he said they’re “great for explaining what a P/E ratio means, walking you through a company’s annual report, or decoding whatever jargon your brokerage app throws at you.”

He also suggested using them to “attack your own ideas.” In other words, you can tell a chatbot why you want to invest in something, then ask it to make the case that you’re wrong—a strategy he deems “worth more than any pick it’ll ever hand you.”

However, it probably isn’t a good idea to directly ask a chatbot, “What stocks should I buy or sell?” Its answer might be right and might be wrong. The problem is that it can be very hard to tell which.

Data from chatbots vs. financial research and trading platforms

For reliable data, financial research and trading platforms using AI (such as Public.com and Investing.com) trump chatbots. Their info comes from real, verified financial databases as opposed to the open internet.

If you get data from a chatbot, always confirm it with trusted, official sources, such as SEC Edgar filings, established financial news outlets, or direct company reports.

Zare said that he frequently sees investors trusting numbers that chatbots have invented. His advice is simple: “If it didn’t pull a live source with a link, don’t trust the number.”

Using chatbots alongside human advisors

Unlike algorithms, a licensed human financial advisor can provide personalized judgment and is legally bound to protect your finances. However, that doesn’t mean they’ll always get everything right, either.

Certified Financial Planner Achim von Bodman of Watter CPA recommends using AI to get better feedback from a human advisor. “Use it to ask your advisor better questions,” he says. “Do not use it instead of a financial advisor.”

Zare agreed that it’s safer to deal with a human advisor who’s a fiduciary, who’ll be “legally on the hook for bad advice.”

Sidestepping the issue by opting for a simple strategy

One of the most commonly cited pieces of financial advice is to keep your investing strategy simple. In practice, this means investing in low-cost, broad-market index funds.

These funds mirror baskets of stocks. A prime example is a fund tracking the S&P 500, an index of the largest and most successful US companies.

A strategy that revolves around investments like this has historically beaten almost every stock-picking robot—and human.5 That’s because any individual stock can go up or down, but in general, the overall market tends to go up over time.

Virtually any LLM will confirm this piece of investing wisdom, and as financial experts like Zare have pointed out, getting that kind of education may be the real advantage of turning to AI tools. Their greatest utility might lie in explaining exactly why most investors don’t need them at all.

Zare summed it up by saying, “AI is not going to give a retail investor an edge over the market. Anyone promising that is selling something. What it can do is make you a more informed and less emotional investor, and for most people, that beats any hot tip.”