66% of people who use AI have already asked it for financial advice, but a new study found major inconsistencies across ChatGPT, Gemini, and other platforms — here’s how to use AI financial advice tools safely.
The Study That Should Change How You Use AI for Money
A new academic study published in the Journal of Financial Planning, authored by finance professors from the University of Georgia and the University of Rome Tor Vergata, tested seven widely available generative AI platforms — ChatGPT, Claude, Copilot, DeepSeek, Gemini, Meta AI, and Perplexity — on identical personal finance prompts. The researchers asked each platform about three common scenarios: how much to keep in an emergency fund, the optimal withdrawal rate from retirement savings, and the ideal composition of an investment portfolio.
The findings, reported by CNBC on July 7, 2026, were not reassuring for anyone treating AI chatbots as a substitute for real financial planning. Researchers found “significant variation” in how the platforms answered the exact same questions, especially around emergency savings recommendations and asset allocation percentages. In other words, ask five different AI tools the same question and you may get five meaningfully different answers — with no obvious way to know which one, if any, is right for your situation.
Why This Matters Right Now
This isn’t a small-sample curiosity. The same reporting cited survey data showing roughly two-thirds of Americans who have used generative AI say they’ve used it for financial advice, and that number climbs to 82% among both Gen Z and millennial users. AI has quietly become a genuine first stop for financial questions for a huge share of the population — which means inconsistent or biased outputs aren’t a theoretical risk, they’re already shaping real financial decisions being made this week.
The researchers also tested something most users never think to check: whether AI recommendations changed when the hypothetical person’s race or gender changed, with everything else about the financial scenario held constant. The study found the platforms’ guidance could shift in ways that weren’t explained by any legitimate financial variable, raising real questions about embedded bias in how these models were trained.
This matters because most people don’t type a single prompt and stop. They ask follow-up questions, and the model’s earlier answers become part of the context for everything that comes after. If the first response leaned toward an overly conservative emergency fund or an unusually aggressive portfolio split, every subsequent answer in that conversation tends to build on that same starting point rather than correcting toward a more balanced view. A biased or simply inconsistent first answer can quietly steer an entire planning conversation in one direction.
What AI Actually Gets Right
To be fair to the technology, the study and prior research on the topic (including a 2024 paper in the Journal of Risk and Financial Management on ChatGPT specifically) found AI chatbots are genuinely useful for high-level financial education. They’re good at explaining why diversification matters, how compound interest works, or why an ETF might be more tax-efficient than a mutual fund in a taxable account. Andrew Lo, director of MIT’s Laboratory for Financial Engineering, has noted that large language models can sound extremely authoritative even when they’re not correct — which is exactly why they work well as a starting point for understanding concepts, but poorly as a source of specific, personalized numbers.
The 4% retirement withdrawal rule is a good example from the study. Most platforms landed close to this widely cited guideline when asked about retirement withdrawals, because it’s a well-documented, frequently repeated rule that shows up constantly in the training data these models learned from. Where the platforms diverged sharply was on more nuanced, personalized questions — exactly the questions where getting it wrong actually costs you money.
How to Actually Use AI for Your Finances
None of this means you should avoid AI tools for money questions entirely — it means using them for the right job. Treat AI chatbots as a research assistant, not a financial planner. Ask them to explain a concept, summarize the tradeoffs between two account types, or help you understand a term you saw in a filing. Don’t ask them to tell you the exact percentage of your portfolio that should be in stocks versus bonds, or the exact dollar amount your specific emergency fund should hold — those answers depend on your income stability, dependents, job security, and risk tolerance in ways a chatbot has no reliable way to know unless you’ve disclosed your entire financial picture, and even then, the study shows outputs vary by platform for reasons that have nothing to do with your situation.
A practical workflow: use AI to generate a first-draft explanation or a list of questions to research further, then verify anything with a specific number attached against a primary source — the IRS for tax and contribution limits, the Federal Reserve for rates, or a fee-only financial advisor for anything that involves your actual account balances and goals. If a chatbot gives you a specific percentage or dollar figure, treat it the way you’d treat a stranger’s opinion at a party: potentially useful as one data point, never as the final word.
It’s also worth running the same question through more than one platform before treating any single answer as settled. If ChatGPT, Gemini, and Claude all land in a similar range on an emergency fund recommendation, that convergence is more meaningful than any one platform’s confident-sounding paragraph. If they diverge significantly, that’s your signal to dig into a primary source rather than pick whichever answer you liked best.
Where AI Tools Genuinely Help Creators and Savers
The picture is different when AI is used for tasks it’s actually built for, rather than for personalized financial planning. AI writing and voice tools can help you build a side income stream, automate repetitive tasks, or produce content faster — none of which requires the model to give you personalized financial advice at all. For creators specifically, AI voice generation tools have become genuinely reliable for producing consistent narration without needing to be on camera, which is a very different use case than asking a chatbot how to allocate a retirement account.
The distinction that matters is between AI as a production tool and AI as a decision-maker. When you use AI to draft a script, generate a voiceover, or research talking points for a video, you remain the one making the final call on what actually gets published — the stakes of a slightly-off AI suggestion are low, and you’re reviewing the output anyway. When you use AI to decide how much of your paycheck goes into a retirement account or which stocks to buy, you’re handing decision-making weight to a tool that this study just showed can give meaningfully different answers depending on which platform you happen to open. The tool itself isn’t the problem; the job you assign it is.
Recommended resources:
If you’re using AI tools to build content or a side income stream rather than for personalized financial advice, ElevenLabs is a solid option for AI voice generation, and vidIQ can help you find what’s actually working in your niche before you script your next video.
Final Thoughts
AI chatbots have become a genuine first stop for financial questions for the majority of people who use them, but this new study is a clear signal that “first stop” should never become “final answer” when real money and specific numbers are involved. Use AI to learn the concepts, ask it to explain the tradeoffs, and let it help you get oriented — then verify anything with a dollar sign or percentage attached against a primary source before you act on it. The technology is a helpful starting point precisely because it’s fast and always available; it’s a risky ending point because, as this study shows, the exact same question can produce meaningfully different answers depending on which platform happens to be open on your phone.
🔗 Related posts: ElevenLabs V1 Model Retirement Checklist | How to Invest in Stocks for Beginners
FTC Disclosure: This post may contain affiliate links. If you click and make a purchase, we may earn a small commission at no extra cost to you.
The information in this post is for educational purposes only and is not personalized financial advice. Always do your own research before making financial decisions.
