One In Four Americans Use AI For Taxes Experts Warn Costly Mistakes
I asked ChatGPT to prepare my tax return, and it confidently provided tax advice without knowing the latest IRS rules or my individual circumstances. A CNBC investigation found that approximately 1 in 4 Americans are now using AI chatbots like ChatGPT and Claude to prepare their tax returns. Users reported: (1) AI recommended incorrect tax deductions that don't exist under current law, (2) AI miscalculated tax brackets, leading to underpayment penalties, (3) AI failed to account for recent IRS rule changes (training lag), and (4) self-employed users were told they didn't need to pay estimated quarterly taxes — an error that can trigger IRS penalties. One entrepreneur asked ChatGPT about deducting a home office and was told 'you can deduct any room you occasionally work in,' which is incorrect under IRS rules requiring exclusive and regular use. Is that right?
1 Answer
As a Certified Financial Planner who has prepared hundreds of tax returns, I can confirm the risks here are very real. The AI giving tax advice suffers from what we call 'training lag' — these models are trained on data that may be months or even years old, while tax codes change annually. In 2025-2026 alone, the IRS updated standard deductions, adjusted tax brackets for inflation, and modified home office deduction rules. The three most dangerous errors I see AI making with taxes: 1. Home office deductions: AI consistently tells people they can deduct 'any room used for work.' This is wrong. The IRS requires exclusive and regular use — a spare bedroom that doubles as guest quarters doesn't qualify. I've had clients come to me after AI told them the wrong thing, facing audits. 2. Estimated quarterly taxes: For self-employed individuals, AI often says 'your employer will handle it' or 'you don't need to worry about it.' Wrong. Self-employed workers earning over $1,000 must pay quarterly estimated taxes or face IRS underpayment penalties. This is one of the most costly AI errors I see. 3. Retirement account contributions: AI frequently miscalculates the interaction between Roth and Traditional IRA contributions, especially when modified adjusted gross income (MAGI) limits come into play. The AI fails to model the income phase-out ranges correctly. The Journal of Financial Planning study (June 2026) tested 7 major AI programs — ChatGPT, Claude, Copilot, DeepSeek, Gemini, Meta AI, and Perplexity — on questions about emergency savings, asset allocation, and retirement withdrawals. All seven showed significant variation in answers, meaning consumers get different (and potentially conflicting) financial advice depending on which AI they happen to ask. My advice: use AI as a starting point for brainstorming tax questions, but always verify against the actual IRS website (irs.gov) or consult a licensed CPA or CFP. The cost of a professional consult is far less than the cost of an audit.
Your answer
Sign in to verify this AI response.
Don't trust us — or the AI. Ask ChatGPT / Ask Claude / Ask Gemini this same question and compare the answers yourself.
More in Finance
- AI Financial Advice 37 Percent Wrong Vanguard Survey — Conversational AI chatbots
- Chatbot Invented Salary 100k Assets Super Advice — Unidentified AI chatbots
- ChatGPT UK Savings Rates Wrong Fscs 85000 Which — ChatGPT (OpenAI)
- Coinbase AI Breaking News Alert Norway Brazil 3 2 Hallucination — Coinbase AI news alert bot
- ChatGPT Claude Deepseek Eur500 Portfolios All Lost Money — ChatGPT, Claude and DeepSeek
- ChatGPT Scam Broker Safety Calls Wrong Two In Three — ChatGPT (OpenAI)
- AI Pension Answers One In Ten Potentially Harmful — Copilot, ChatGPT, Gemini and Claude
- ChatGPT Gemini Cac40 Half Year Figures Wrong — ChatGPT and Gemini
- Revolut Rita Chatbot Break Fee Already Waived — Revolut chatbot Rita
- AI Chatbots Wrong Tax Answers Nerdwallet — ChatGPT, Gemini and Perplexity
More from this topic
I asked a chatbot how to split my retirement savings between stocks and bonds. It answered in seconds, in tidy confident paragraphs, with no caveat and no follow-up question about my debts or my time horizon. I assumed it was right. Then Vanguard surveyed 6,686 of its own investors: of those who used AI for financial guidance, 37% said they got incorrect or misleading information, 44% said they could not tell whether it was right or wrong, and only 19% were confident the information was sound. Is that right?
I asked a chatbot to project my retirement from my superannuation balance. It came back with a complete, self-assured set of numbers. Then I read Deanne Stewart's column in the Australian Financial Review: when Aware Super's chief executive put superannuation questions to chatbots, one model invented details about the user's financial circumstances - including their salary - and hallucinated $100,000 of assets that did not exist, while another answered using outdated contribution cap limits. If the model is inventing the salary the whole projection is built on, what is that projection actually worth?
I asked ChatGPT: 'What's the best savings rate?' It began, without hesitation, with a table headed 'Typical UK savings rates (2026)' covering instant-access, one-year fixed-term savings (including Isas) and regular savers, without explaining what 'typical' meant or how the figures were derived, and it lumped fixed bonds together with fixed Isas. It quoted: instant-access average 2.15% AER; one-year fixed-term average 3.85% AER; one-year fixed-term cash Isa average 3.85% AER; regular savers up to 7.5% AER. For top rates it named 4% to 4.5% AER on instant-access accounts, up to 4.39% AER on instant-access cash Isas and 4.1% to 4.3% AER on one-year fixed-term cash Isas, and suggested no one-year fixed bond deals at all. Asked where to put a GBP 25,000 lump sum, it said the top 4.5% AER rate was available from Chase or Marcus by Goldman Sachs. Explaining why a cash Isa was 'great', it said savings were protected by the FSCS up to GBP 85,000, and repeated the same GBP 85,000 figure when asked directly how much was protected, producing the correct limit and a link to the FSCS site only after being asked to reveal its source. Is that right?