Does AI Give Women Worse Financial Advice Than Men?

Here's the answer up front: yes. A new MIT Sloan and Stanford Graduate School of Business study found that AI chatbots recommend safer, lower-return investments to women than to men, even when the two people are asking almost the exact same question. Run that gap forward to retirement and it adds up to roughly $59,890 less in savings for a woman who followed the same "sound, cautious" advice a man got from the identical bot.

What did the MIT and Stanford researchers find?

Economists Tim de Silva (Stanford GSB), Taha Choukhmane, Weidong Lin, and Matthew Akuzawa (MIT Sloan) built a working paper called "AI Financial Advice: Supply, Demand, and Life Cycle Implications," where they simulated a saver's entire working life on the advice of GPT-5.2 and Gemini 3 Flash. The topline finding: AI recommends about three percentage points less stock exposure to women than to men, which compounds into a 4.1% smaller nest egg by age 60.

Two-thirds of that gap traced back to how people phrased the question. Prompts from men leaned on words like "portfolio," "growth," and "strategy." Prompts from women leaned on "family," "grocery," and "credit." The models picked up on that more circumstance-focused language and answered back with more cash, fewer stocks, less risk, less growth. But the last third is harder to explain away as a prompting quirk. When the researchers kept a question word-for-word identical and changed only the stated gender of the person asking it, the model still recommended less equity to the woman. That's not her doing anything differently. That's the model deciding, on its own, that a woman can carry less risk than a man asking the identical question.

The pattern you've seen before

If this rhymes with things you've read here before, it should. Pew's own numbers show women already use AI less than men do, which means fewer women are even in the room to notice a gap like this exists in the first place. It's not just financial advice, either — women report getting judged more harshly than men for the exact same AI use at work. And when you factor in how few women are actually leading the research that builds these models, a $60,000 retirement gap starts to look less like a fluke and more like a pattern nobody caught, because the people stress-testing these systems weren't the ones typing "grocery" and "family" into the prompt box.

Why this matters more if you're younger

If you're 26 and just opened your first Roth IRA, or you're the one in your friend group who actually reads the fine print, this is exactly the decade where a few percentage points of stock exposure compounds into real money. A 28-year-old steered toward "safer" investments today isn't losing a little now — she's losing four decades of growth on the difference between what she was told and what a man asking the same question was told. And because more of us are turning to AI first, before a human advisor, before a parent, before even a basic Google search, the advice we get in that very first conversation is doing more quiet, structural work than it looks like it's doing.

One thing to try today: the next time you ask an AI chatbot a money question, write the prompt like you're stating what you want, not just describing your life. Instead of "I have some savings and a family, what should I do," try "I'm 28, I want to maximize long-term growth, and I can tolerate short-term volatility — what asset allocation would a fee-only fiduciary recommend?" The researchers found that specificity alone closes a real chunk of the gap. It costs you ten extra seconds of typing, and it's worth it.

Quote to sit with

"A woman must have money and a room of her own if she is to write fiction." — Virginia Woolf

💌 Sarah