Ask a chatbot where to put your retirement savings, and you will get an answer in seconds. That speed explains why half of Americans have turned to artificial intelligence for financial guidance. Young people lead the charge, with 82% of Millennials and Gen Z using AI to manage their money, according to a new survey. The numbers sound impressive until you look at what the research actually found.
A working paper from the MIT Sloan School of Management ran detailed simulations on AI financial advice. The results were mixed enough to make anyone pause before acting on a chatbot's recommendations.
AI Gets Some Big Things Right
Tim / Pexels / The MIT researchers discovered that AI pushes people toward smart money habits. The chatbots recommend saving more during your working years and spending less in retirement.
They tell you to put money into diversified stock funds instead of chasing hot tips. And they suggest reducing your investment risk as you get older.
Following AI's general advice in the MIT simulations led to a portfolio worth over $1,000,000 by retirement age. That is real money for regular people who cannot afford a human financial planner. The technology also forces a broader conversation about money.
Only 6% of users mentioned liquidity in their questions, but AI brought it up in 83% of its answers. It introduced saving strategies and safer investments even when people did not ask for those topics directly.
Your Chatbot is Confidently Wrong
A separate audit of ChatGPT, Perplexity, Gemini, and Claude found something disturbing. The financial advice from these platforms is frequently outdated, incomplete, or just plain inaccurate. These are not small mistakes either. The audit showed that AI engines kept warning users about a reduction in the federal estate tax exemption.
That warning became obsolete in July 2025 when new legislation permanently raised the exemption.
The errors follow a pattern. AI loves to talk about the benefits of complex strategies like premium financing. But it buries the risks, things like interest rate exposure and collateral calls. The chatbots sound so confident that people trust them.
Prior research found hallucination rates between 20% and 37% for financial citations generated by major AI models. That means up to one-third of the sources your chatbot cites do not actually exist.
The Gender Gap You Did Not Expect
Nilov / Pexels / The advice generated from prompts written by women led to nearly $60,000 less wealth compared to advice from prompts written by men.
That gap did not happen because women know less about money. It happened because of how the AI interprets language.
Women tend to use words like “family” and “grocery” in their prompts. Men use words like strategy and growth. The AI picks up on those signals and changes its recommendations accordingly. Even when researchers sent identical prompts labeled as coming from a woman versus a man, the AI recommended less stock exposure for the woman.
The machine learned gender bias from the data it was trained on, and now it passes that bias back to you.
Your financial literacy level also matters enormously. People with low financial literacy who followed AI advice ended up with nearly $50,000 less wealth by age 60 compared to high literacy users. The same income, the same returns, just different prompts. The chatbot does not know what you do not know, and it will not ask clarifying questions the way a human advisor would.
When a human financial advisor gives you bad advice, you can sue them. You can report them to regulators. They can lose their license or face criminal charges. That pressure keeps human advisors reasonably honest and careful.
