IRS Clarifies How Circular 230 Rules Apply to Tax Pros Using AI

The IRS Office of Professional Responsibility has issued detailed guidance clarifying that tax practitioners using artificial intelligence tools remain bound by all existing Treasury Circular 230 requirements, with no exceptions for AI-assisted work.

The bulletin, posted in late June 2026, establishes that practitioners bear complete responsibility for any output generated with AI assistance. The core message: AI must augment professional judgment, never replace it.

Full accountability for AI outputs

Practitioners must exercise due diligence in verifying the accuracy of all facts, citations, and calculations that AI systems produce, according to the guidance. The IRS characterized AI-generated content as a starting point requiring human scrutiny and editing, not a finished product.

The agency warned that practitioners cannot rely solely on AI outputs and must understand both tax law and the AI tools they deploy, including how systems generate content and where errors or bias may emerge. The bulletin stated that lack of technological competence could lead to improper advice or flawed filings.

Billing and fee considerations

The guidance addresses a practical concern for firms: how to bill for AI-assisted work. The OPR warned that charging for time not actually spent due to AI efficiencies could raise concerns about unconscionable fees. Cost savings from AI should be reflected in client billing, and practitioners should disclose AI use appropriately while fairly crediting clients for any cost reductions.

Firm leaders face obligations to implement compliance procedures covering staff training, secure data handling, accuracy monitoring, and vetting of third-party AI tools, with documentation demonstrating adherence to Circular 230.

Data security and penalties

The bulletin emphasized risks surrounding taxpayer data, pointing to Code provisions imposing civil and criminal penalties for unauthorized disclosure of tax return information. Generative AI platforms, particularly public or unsecured ones, may expose confidential data.

Practitioners must handle all client data using only secure, enterprise-approved AI systems, the agency stated. Willful mishandling of information could trigger disciplinary action under Circular 230.

Why it matters

As tax professionals increasingly adopt generative AI for research, document preparation, and analysis, the IRS is establishing clear expectations that existing professional standards apply without modification. The guidance signals that regulators will hold practitioners accountable for AI errors just as they would for human mistakes, potentially exposing firms to sanctions if they treat AI outputs as reliable without verification. This creates compliance and training obligations for any tax practice using these tools.

Lessons from legal sanctions

The OPR pointed to real-world consequences in the legal field, where courts have sanctioned lawyers for filings containing fabricated citations and other AI-generated errors. Penalties have included financial sanctions, public censure, and removal from cases. The bulletin noted that while most cases have involved attorneys, similar risks apply to tax professionals, citing an example where an accounting firm's report for the Australian government included invented quotes and incorrect citations, prompting a partial fee refund.

The guidance was first reported by the Journal of Accountancy. The IRS Office of Professional Responsibility enforces compliance with Circular 230 and can discipline practitioners for violations.