August 1, 2026
Wall Street meets chatbot chaos
From MIT: AI financial advice is surprisingly good
MIT says AI money tips work — but commenters say the fine print is the real boss
TLDR: MIT found that AI chatbots can give pretty solid money advice and may help many adults save more, especially when people ask detailed questions. Commenters were split between “wow, useful” and “of course it works only with perfect prompts,” with plenty of jokes that humans give worse advice anyway.
A new MIT study basically walked into the internet and said, "guess what, the robot might actually be better with your money than you are." Researchers found that when people followed advice from popular AI chatbots, they generally saved more, invested more sensibly, and built healthier cushions for later life — especially after age 30. The bots tended to push classic good habits: save during working years, spend down in retirement, and avoid going too wild with risk as you get older. But the comments section was not about to let the machines have an easy victory lap.
The loudest reaction was a giant asterisk: AI is only “surprisingly good” if you ask it the right way. One commenter roasted the headline itself, saying the missing half was doing “heavy lifting.” Translation for non-finance readers: if you give the bot clear details about your age, income, savings, and goals, the advice gets much better. Shocking! Other commenters argued the bots are still too bland and overly cautious, warning that AI often defaults to safe, mainstream answers and may dismiss more advanced investing strategies that some people swear by.
Then came the jokes. One user said AI works great until you ask it to justify your terrible choices, which is probably the most relatable money advice on the entire internet. Another cut straight to the knife with: human financial advice is surprisingly bad. And one practical commenter brought receipts, saying they exported their budgeting data from YNAB and got genuinely useful help from Claude. So yes, the robots may be decent at money — but the real drama is whether they’re wise guides, boring scolds, or just the newest enablers for your financial delusions.
Key Points
- •MIT researchers found that financial advice from large language models was generally strong and often aligned with established financial principles.
- •The study used a life-cycle benchmark model and simulated outcomes for people aged 22 to 89 who repeatedly followed AI advice on spending, saving, and investing.
- •Participants’ prompts to GPT-5.2, GPT 5.6, and Gemini 3 Flash were compared with more detailed academic prompts containing full financial information and economic assumptions.
- •Following AI recommendations could create sizable saving buffers for virtually all individuals over age 30, according to the article.
- •The models were weaker at handling shocks such as unemployment and often failed to rebalance portfolios actively, allowing drift over time.