CONTINUING EDUCATION FOR TAX & FINANCIAL PROFESSIONALS

The 2026 Summer Federal Tax Update is here – Get ahead of client questions with mid-year insights

IRS Issues Its Introductory Guidelines on Responsible AI Use

Western CPE Staff

On June 24, 2026, the IRS Office of Professional Responsibility (OPR) released Alert 2026-19, Introductory Guidelines for Responsible AI Use in Federal Tax Practice. It is the agency’s introductory guidance on how artificial intelligence fits within the ethical duties tax professionals already carry. Every obligation you hold under Treasury Circular 230 still applies when you hand part of your work to a machine.

The OPR frames AI as a tool that supports professional judgment rather than one that replaces it. Responsibility for the final product of your work stays with the qualified professional, no matter how that work was produced.

Rising Use of AI and Its Risks

Most tax and accounting firms rely on AI in some form, through research platforms and document review tools. Newer generative systems go a step further by creating original content largely on their own. Those capabilities create new risks. The OPR points to three well-documented weaknesses of generative AI:

Fabricated outputs, often called hallucinations, where the system invents facts, cases, or citations.
Bias built into the underlying data and models.
A lack of transparency, meaning users cannot always see how a conclusion was reached.

The OPR also flags a privacy problem specific to shared AI systems. Information entered for one client can resurface in an answer prepared for another. That kind of data bleed can compromise confidentiality even when no one intends any harm.

Circular 230 Duties and AI

The OPR maps AI use onto Circular 230 provisions already on the books:

Due diligence (§10.22). Practitioners must verify the facts, figures, and citations AI produces before anything reaches a client or the IRS.
Fees (§10.27(a)). Billing for hours that AI actually saved can cross into an unconscionable fee. The OPR expects efficiencies to be reflected in billing and disclosed as appropriate.
Competence (§10.35). Competence now includes understanding how your AI tools work, where they fail, and when their output is unsuitable for a filing.
Firm procedures (§10.36). Firm leaders must put policies in place for training, data handling, and tool oversight (more on this below).
Written advice (§10.37). Advice must rest on reasonable assumptions. If a system’s reasoning cannot be examined, leaning on it may itself be unreasonable.

Confidentiality (IRC §§6713 and 7216). Civil and criminal penalties apply to unauthorized disclosure of return information, a real danger when data is uploaded to public tools.

Unchecked AI Hallucinations Have Consequences

The OPR stresses that these problems don’t just apply to law. It cites a 2025 report prepared for the Australian government by Deloitte Australia that contained fabricated quotes and references to sources that did not exist, prompting the firm to refund part of its fee. For tax professionals, the same kind of error carries the same kind of risk.

Firm Leader Obligations

The obligations under §10.36 fall on the people who run a practice. According to the guidance, firms should build and document procedures that cover:

Staff training on both the capabilities and the risks of AI.
Secure protocols for handling client data and monitoring accuracy.
Vetting of any third-party or outsourced AI tools before they are used.

Leaders who skip these steps can face discipline if a pattern of noncompliance takes hold across the firm.

Stay Up to Date on State Laws and Regulations

The OPR reminds practitioners that federal ethics rules are not the only standard in play. Several states, including California, Colorado, Illinois, and Utah, have enacted laws addressing AI transparency, bias, and consumer protection. Professional bodies have weighed in as well, with the American Bar Association issuing a formal ethics opinion on generative AI in 2024. Keeping current on these state laws and professional standards, alongside the federal rules, is part of practicing responsibly.

Guidance for Practitioners

The OPR alert recommends some practical habits to follow:

Track the federal and state rules that apply to your work.
Secure client data and control who can access it.
Keep records of how AI is used and how its output is verified.
Never upload sensitive information to unsecured platforms.
Treat anything AI writes as a first draft and review it in full for accuracy and bias.

Be Responsible With AI In Your Practice

The OPR’s alert is only introductory guidance on the use of AI. As the technology advances and the rules around it shift, the agency expects practitioners to keep up with evolving guidance from government agencies and professional bodies. AI can make tax work faster and sharper, but practitioners remain accountable. Judgment and decision-making belong to the professionals.

You can read more about the the IRS’s introductory guidelines for AI usage here: