As tax professionals prepare for the upcoming filing season, the IRS Office of Professional Responsibility (“OPR”) has issued its first formal guidance addressing the use of artificial intelligence by tax practitioners. OPR Alert 2026-19, released June 24, 2026, explains how existing professional standards under Circular 230 apply to artificial intelligence (“AI”) and generative artificial intelligence (“GAI”). 

The Alert does not amend Circular 230 or create new professional-responsibility rules. Instead, it confirms that practitioners remain responsible for work performed with AI assistance, including the accuracy of tax filings and advice, the protection of client information, and the reasonableness of fees. 

What Is New

The Alert is the OPR’s first formal interpretation of Circular 230 in the context of AI. It expressly connects the use of AI to practitioners’ existing obligations concerning due diligence, competence, written advice, fees, firm procedures, and the confidentiality of tax return information. The central message is clear: AI may assist a practitioner, but it cannot replace the practitioner’s professional judgment or serve as the practitioner of record. Practitioners must independently review and verify AI-generated work before submitting it to the IRS or providing it to a client.

Independent Review of AI Output

Circular 230 requires practitioners to exercise due diligence under Section 10.22 and possess the competence necessary to handle a matter under Section 10.35. In the AI context, these duties require practitioners to understand the capabilities, risks, and limitations of the tools they use. Practitioners should treat AI-generated material as a preliminary draft. They must independently verify factual statements, calculations, legal authorities, citations, projections, and other substantive content.

For example, if a client asks a practitioner to prepare a nexus study for state income and sales tax purposes, the practitioner may use AI to help produce the study but the practitioner must independently verify all factors that are used in the study: sales thresholds, transaction thresholds, tax rates and effective dates of all laws on a state-by-state basis.

Written Tax Advice

When practitioners use AI to draft written advice concerning federal tax matters, Section 10.37 continues to require reasonable factual and legal assumptions and consideration of all relevant facts and circumstances. Practitioners should not accept AI-generated citations, forecasts, calculations, or conclusions at face value. They must authenticate the underlying information and determine whether the resulting advice is reasonable. The Alert also warns that reliance on an AI system may be unreasonable when the system’s underlying logic is opaque and the practitioner cannot meaningfully evaluate its output.

For a tax practitioner, AI can be a valuable tool in producing a tax opinion letter.  But before the opinion is finalized, the responsible practitioner must verify the legal analysis, confirm the transaction facts, and replace any fabricated, outdated, or inapplicable authorities before finalizing and delivering the opinion.

Client Privacy and Tax Return Information

The Alert emphasizes the risks of entering client information into public, unsecured, or unapproved AI platforms. Unauthorized use or disclosure of tax return information may lead to civil or criminal penalties under Internal Revenue Code Sections 6713 and 7216.

Before using an AI tool, practitioners and firms should determine:

• Whether the tool has been approved for the intended use;
• What information the tool collects and retains;
• Whether user inputs may be used to train the provider’s models;
• Who may access information submitted to the tool; and
• Whether appropriate confidentiality, security, and access controls are in place.

Practitioners should use only secure, firm-approved AI systems when handling client information and should never upload tax return information or other sensitive data to a public or unsecured AI platform. Consider, for example, a practitioner who pastes a client’s tax data and financial information into a consumer AI chatbot to obtain help with a return or analysis. Even if the prompt omits the client’s name, the combined figures and details may identify the taxpayer and constitute protected tax return information, so the practitioner should use only approved, secure systems.

Fees for AI-Assisted Work

Circular 230 Section 10.27 prohibits practitioners from charging unconscionable fees. The Alert cautions that billing practices should accurately reflect efficiencies created by AI. For example, a practitioner should not bill a client for hours of manual research or drafting that the practitioner did not perform because an AI tool completed the work more quickly. Firms should review their billing practices to ensure that invoices accurately reflect the work performed and appropriately account for AI-related time savings.

Firm Policies and Training

Section 10.36 requires firm leadership to maintain adequate procedures for compliance with Circular 230. According to the Alert, this responsibility includes establishing policies governing the use of AI. Before the upcoming filing season, firms should:

• Identify which AI tools employees may use and for what purposes;
• Prohibit the entry of client information into public or unapproved platforms;
• Train personnel on AI risks, confidentiality requirements, and verification obligations;
• Establish review procedures for AI-assisted filings and written advice;
• Document AI use and the steps taken to verify resulting work;
• Evaluate third-party AI providers before permitting their use; and
• Monitor AI tools for accuracy, security, and compliance.
• Firms should also remain aware of applicable federal and state AI requirements.

Preparing for the Upcoming Tax Season

Tax practitioners do not need to stop using AI. They do, however, need to use it within the same professional-responsibility framework that applies to all other aspects of tax practice. Before filing season begins, practitioners and firms should confirm that their AI practices provide meaningful human review, protect client information, produce accurate and supportable work, and comply with documented internal procedures. The practitioner remains responsible for the final filing or advice, regardless of how much of the underlying work an AI system performed.

For questions about the use of AI in tax practice, Circular 230 compliance, or developing AI governance policies for the upcoming filing season, please contact David Hughes another member of Kilpatrick’s Employee Benefits & Tax team.