The Intersection of AI and Tax Professional Standards

The integration of Artificial Intelligence (AI) in tax preparation and research has accelerated significantly, prompting the Internal Revenue Service (IRS) to issue new warnings regarding professional responsibilities. In recent guidance, the IRS Office of Professional Responsibility (OPR) explicitly outlined how Circular 230 duties apply to practitioners utilizing AI tools in their daily practice.

While AI promises unprecedented efficiency in tax research, drafting, and document processing, it also introduces substantial risksβ€”ranging from data privacy breaches to algorithmic "hallucinations." The IRS has made it clear that tax professionals cannot delegate their due diligence obligations to software, no matter how advanced it may be.

The Core Message: Human Oversight is Non-Negotiable

Circular 230 establishes the rules governing practice before the IRS. A central tenet of these regulations is the practitioner's obligation to exercise due diligence as to accuracy. According to the latest OPR guidance, a practitioner who relies on an AI tool to prepare a return, draft a legal opinion, or respond to an IRS notice remains fully and personally liable for the accuracy of that work.

"Artificial Intelligence is a tool, not a practitioner," the IRS emphasized. The responsibility for evaluating the accuracy and legal soundness of any position always rests with the human professional signing the return or representing the taxpayer.

Key Risks Highlighted by the IRS

The recent guidance focused on several specific risks associated with generative AI and automated tax software:

  • Hallucinations and Inaccurate Citations: AI models have been known to invent case law or misinterpret complex sections of the Internal Revenue Code. Practitioners must verify every citation, calculation, and legal conclusion generated by AI against authoritative primary sources.
  • Client Confidentiality: Inputting sensitive taxpayer information (such as Social Security numbers, financial data, or trade secrets) into public or unsecured AI models can constitute a direct violation of Internal Revenue Code Section 7216 and Circular 230 confidentiality rules.
  • Lack of Competence: Using a sophisticated tool without understanding how it derives its conclusions can violate the fundamental duty of competence. Practitioners must understand both the capabilities and the limitations of the technology they employ.

Best Practices for Tax Professionals

To remain compliant while leveraging new technology, the IRS suggests that firms adopt strict internal policies regarding AI use. This includes:

  • Using Enterprise-Grade Tools: Utilizing only secure, closed-environment AI tools that guarantee data privacy and explicitly state they do not use client inputs to train public models.
  • Mandatory Human Review: Implementing a strict human review process for all AI-generated work product before it is finalized, signed, or submitted to the IRS.
  • Staff Training: Educating all firm employees on the ethical limitations and potential biases of AI tax software.

The Bottom Line

As the tax profession continues to evolve, the message from the IRS is clear: technological innovation is welcome, but professional accountability cannot be automated. Practitioners who embrace AI without implementing adequate safeguards risk facing severe disciplinary action from the OPR, monetary penalties, and permanent damage to their professional reputation.