Tax Preparers Adopt AI, but Privacy Rules Lag Behind

Tax professionals are integrating artificial intelligence into their practices at an accelerating pace, but federal privacy protections designed for an earlier era may not adequately cover how client data flows through these new systems.

The IRS issued its first AI guidance for tax practitioners in June, requiring professionals to review and verify AI-generated work and adjust billing to reflect efficiency gains. Notably absent: any requirement to disclose to clients when AI is used to prepare their returns.

Existing law—Section 7216 of the Internal Revenue Code—governs when tax preparers must obtain signed client permission before sharing information with third parties, such as financial advisors. But the statute includes exceptions that have traditionally applied to tax software, and that's where ambiguity now exists.

"The problem with 7216 is that the last formal guidance we've received from the IRS goes back to 2013," said Henry Grzes, lead manager for tax practice and ethics at the American Institute of Certified Public Accountants. "When you think about how the tax practice landscape has changed in those 13 years, we're really hoping for some additional guidance."

Rapid adoption across the profession

A June survey from Blue J and CPA.com found that 60 percent of tax professionals now use AI for tax research at least weekly, up from 33 percent in 2025. The survey of more than 1,000 practitioners also revealed that 44 percent use AI for advisory projects, 40 percent for tax planning, and 36 percent for document analysis.

An earlier 2024 Thomson Reuters Institute report showed about 25 percent of tax and accounting professionals using public-facing generative AI tools, with just 9 percent using proprietary tax-specific AI technology.

Why it matters

The gap between technology adoption and regulatory clarity creates risk for both practitioners and clients. Tax preparers face potential penalties—including fines up to $1,000 or jail time—for knowingly violating privacy statutes. Clients, meanwhile, may have their sensitive financial data processed by AI systems without their knowledge or consent, and the safeguards protecting that information vary widely depending on the platform.

Not all AI tools are equal

Joshua Youngblood, founder of The Youngblood Group in Dallas and an IRS enrolled agent, argues that AI tools differ fundamentally from traditional tax software. "A lot of people are relying on AI to make judgment calls and to tell them what to do," he said. "I would argue that [an AI tool] is really not just like your tax software."

Grzes noted a critical distinction: some AI platforms use client input to train their models, while closed systems do not. That difference could be addressed in future IRS guidance.

Elizabeth Beastrom, president of tax, audit and accounting professionals at Thomson Reuters, said in an email that when AI platforms use client data solely to prepare that specific client's return, with data protected and used only for that purpose, the tool functions similarly to other software practitioners already rely on—with the practitioner remaining fully accountable.

What clients should ask

Grzes recommends tax practitioners obtain signed disclosures from clients when using AI to prepare returns, even in the absence of formal guidance. "Be safe, as opposed to finding out, 'uh-oh, I should have gotten this,'" he said.

For consumers, he suggests asking directly whether AI is involved and what safeguards are in place. "If the response is something along the lines of 'don't worry,' 'everything is safe' or something similar, without explaining what those safeguards are, I would be concerned," Grzes said.

The AICPA has asked the IRS to prioritize guidance on technology use—including AI—in tax preparation as part of its 2026 recommendations.

These details were first reported by CNBC.