Tax Experts Warn Against Using AI Chatbots for Tax Preparation Despite Industry Push

2 Sources

Share

As tax season approaches, xAI executives are encouraging users to have Grok check their taxes, claiming significant refund increases. But tax professionals and industry experts are pushing back hard, warning that general-purpose AI chatbots pose serious accuracy and privacy risks. Tests show chatbots miscalculate refunds by an average of over $2,000, while data security concerns loom large.

xAI Pushes Grok for Tax Returns, Sparking Expert Backlash

James Burnham, xAI's General Counsel, recently urged his followers to use Grok to double-check tax returns, claiming a friend increased her tax refund by $1,400 through the AI chatbot

2

. The pitch comes as the April 15 filing deadline approaches and taxpayers seek shortcuts through complicated federal income tax returns. Yet tax professionals are issuing stark warnings about using AI and taxes together, particularly when it comes to general-purpose AI chatbots like Grok, ChatGPT, Claude, Gemini, or Perplexity for actual tax preparation.

Source: Gizmodo

Source: Gizmodo

"I don't recommend that at all," said Travis Thompson, a tax attorney and director in the business and finance group at Fennemore

1

. Sterling Raskie, senior lecturer of finance at the University of Illinois Urbana-Champaign, echoed this sentiment: "My advice would be no"

1

. The Internal Revenue Service expects financial data to be accurate, not just "close enough," creating a fundamental mismatch with how Large Language Models operate.

AI Inaccuracy Creates Serious Financial Liability

The core problem with using AI for tax preparation lies in how these systems function. At their foundation, AI chatbots are pattern-finding machines that offer plausible results but cannot distinguish approximation from truth

1

. When The New York Times tested multiple chatbots using training materials from TurboTax partner TaxSlayer, they found the AI miscalculated refunds and amounts owed to the IRS by an average of more than $2,000

2

.

Source: CNET

Source: CNET

TaxCalcBench, a benchmarking test designed to evaluate an AI model's ability to calculate tax returns, revealed that most options fail to even reach 50% accuracy across a full return

2

. This means Grok might show a $1,400 refund increase not because it discovered hidden exemptions, but simply because it entered numbers incorrectly or fabricated deductions.

"It's important to keep in mind that if an AI chatbot provides incorrect guidance and a person uses it to file an incorrect tax return, they (the person) are responsible for infractions or violations, which could include penalties, interest, and lost refunds," said Chris Linderwell, vice president of consumer tax products at H&R Block

1

. This user accountability means any AI tax advice that leads to errors becomes your legal problem, potentially resulting in audit risk, expensive repayments, and fines.

Privacy Risks With AI and Sensitive Financial Data

Beyond accuracy concerns, tax expert warnings emphasize the dangers of sharing sensitive financial data with AI chatbots. Highly personal information like Social Security numbers, W-2 forms, and financial statements should be kept secure, yet chatbots manage data in the cloud on servers owned by private companies

1

.

"You don't want those numbers floating around the internet," Thompson warned

1

. Joel Salas, owner of Elevated Tax Strategies, told Gizmodo: "Honestly, it hasn't been around long enough for us to trust these companies with that type of data. If you review any of their terms and conditions and fully understand them, you'll realize you're playing a bit of Russian roulette with your data"

2

.

Researchers at Stanford found that many leading US AI companies, including Anthropic, OpenAI, and xAI, feed user inputs back into their models unless users manually opt out

2

. Grok itself was recently found revealing people's personal information to other users, while Meta previously allowed users to browse other people's prompts, exposing medical, legal, and other sensitive information

2

. Data security remains a critical concern as the Musk-led Department of Government Efficiency reportedly sought access to sensitive taxpayer data at the IRS

2

.

Limited But Safe Uses for AI During Tax Season

While you shouldn't trust AI chatbots to fill out tax forms, they can serve as educational tools when used carefully. An OpenAI representative confirmed that ChatGPT cannot access bank accounts nor act as a licensed financial professional, lawyer, or accountant, adding "You should always review the ChatGPT output since it is not a replacement for a licensed professional"

1

.

AI can help translate tax code terminology, prepare checklists, or generate questions to ask your accountant. It might help identify tax situations you hadn't considered, like implications of cryptocurrency income or differences between filing statuses. However, you must fact-check any information against reputable sources, and remember that answer quality depends heavily on how you phrase questions

1

.

"Tax preparation in particular requires purpose-built systems designed for accuracy, compliance, reliability, and security," an Intuit spokesperson explained

2

. While H&R Block and other tax giants have introduced AI assistants for tax-related questions, these differ significantly from general-purpose chatbots. Salas pointed to specialized tools like Taxbox as better alternatives for those seeking AI assistance as a "tax assistant or tutor"

2

.

Thompson offered a final caution: "The less and less expertise you have in that field, the less and less you can trust those programs"

1

. As AI companies push their products into tax season, taxpayers should watch for clearer guidance from the IRS on AI use and consider whether purpose-built tax software or human professionals better serve their needs.

Today's Top Stories

TheOutpost.ai

Your Daily Dose of Curated AI News

Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Spend less time searching for the latest in AI and get straight to action.

© 2026 Triveous Technologies Private Limited
Instagram logo
LinkedIn logo