Lappu AI Lappu AI Get Started Free

← Back to Blog

Lappu AI Journal

Responsible AI Use in Tax and Accounting: Supervision, Assessment, and Governance

AI is not automatically forbidden in tax and accounting practices, but firms need competent supervision, human review, client data protections, and clear governance before adopting AI-assisted workflows.

Vivek Uppal • May 23, 2026

Responsible AI Use in Tax and Accounting: Supervision, Assessment, and Governance

Artificial Intelligence is becoming part of everyday business operations, including within the tax and accounting industry. Firms are evaluating AI for email drafting, document summarization, workflow automation, research assistance, internal training, and operational efficiency. That evaluation is appropriate. AI is not something the profession can simply ignore. At the same time, some industry publications have used very strong language suggesting that AI tools are categorically prohibited under regulations such as IRS Circular 230 or incompatible with FTC Safeguards Rule obligations.

That framing is too broad.

The responsible position is not that AI is automatically forbidden. The responsible position is that AI must be used carefully, competently, and under appropriate professional supervision. Tax and accounting firms should not adopt AI blindly. They should also not be pushed into fear-based conclusions that are not supported by the actual language of the relevant rules.

Circular 230 Does Not Ban AI

IRS Circular 230 establishes standards governing practice before the IRS. It focuses on professional responsibilities such as due diligence, competence, confidentiality, accuracy, practitioner responsibility, supervision, and standards for written advice.

Circular 230 does not explicitly prohibit the use of Artificial Intelligence tools.

It does not single out a particular AI platform, provider, or technology category as impermissible. The principles within Circular 230 apply broadly to how tax professionals perform their work, whether the work involves spreadsheets, calculators, cloud software, outsourced services, junior staff, or AI-assisted systems.

That is the more useful way to think about AI in the tax and accounting profession.

AI does not remove professional responsibility. It makes governance, review, and judgment even more important.

This Is About AI Generally, Not One Specific Vendor

Discussions around AI in accounting often become overly focused on a specific provider or product. That framing misses the broader issue. The important question is not whether one individual AI platform is “good” or “bad.” The important question is whether the firm using AI has implemented appropriate safeguards, review processes, and professional oversight. Any AI system, regardless of vendor, introduces considerations around data confidentiality, accuracy, retention, human review, vendor risk, internal controls, and client communication standards.

A firm should evaluate AI systems based on how they are used, what data is shared with them, what contractual protections exist, whether outputs are reviewed, and whether the firm has a clear policy governing their use.

The Real Risk Is Unsupervised Reliance

One of the most important principles within Circular 230 is the obligation to exercise diligence and competence. That obligation does not disappear simply because AI is involved. If a practitioner uses AI-generated output without verification, review, or professional judgment, significant problems can arise. AI systems can produce inaccurate statements, incomplete analysis, fabricated citations, outdated information, or misleading conclusions presented in a confident manner.

For that reason, AI-generated work product should never be treated as automatically authoritative.

Tax professionals remain responsible for verifying facts, reviewing conclusions, confirming the applicability of tax law, exercising independent professional judgment, and ensuring that advice provided to clients is accurate and appropriate.

This is where the real Circular 230 issue lives.

The risk is not the mere existence of AI in a firm’s workflow. The risk is unsupervised reliance on AI-generated output in situations where professional diligence, competence, and judgment are required.

Our Approach: Human Review and Client Data Protection

At our company, we follow two simple principles:

  • Human Review: We never provide clients with AI-generated advice unless it has been reviewed by a human.
  • Data Privacy: We do not share sensitive client data with AI systems or platforms.

These principles reflect the way professional services should approach AI.

AI can assist with drafting, organizing, summarizing, and researching. But it should not replace accountability. A human professional must remain responsible for what is ultimately communicated to a client.

That is especially important in tax and accounting, where clients may rely on professional advice for financial, legal, regulatory, and compliance decisions.

AI Governance Is Becoming Part of Responsible Practice

As AI becomes more common, firms need to move beyond informal experimentation.

They need AI governance.

AI governance does not have to be overly complicated, especially for small and mid-sized firms. But firms should have a clear internal understanding of what AI tools are approved, what data may not be entered into AI systems, which use cases require human review, and who is responsible for approving AI-assisted work product.

The purpose of AI governance is not to block innovation. The purpose is to allow firms to use AI responsibly while protecting clients, preserving professional judgment, and meeting regulatory expectations.

This is where the profession should be heading: not toward blanket fear of AI, and not toward careless adoption, but toward governed, supervised, and accountable use.

FTC Safeguards Rule Considerations

The FTC Safeguards Rule introduces additional responsibilities around the protection of customer information. Tax and accounting firms routinely handle highly sensitive data, including Social Security numbers, tax returns, payroll information, banking details, identity documents, and financial records.

When firms use AI systems, they must evaluate how data is transmitted, whether prompts are retained, whether information is used for model training, what vendor security controls exist, and how access, retention, encryption, and storage are handled.

These concerns are legitimate and should be taken seriously.

But again, this does not mean AI itself is prohibited. It means firms must apply appropriate security and governance controls when incorporating AI into business operations.

A tax or accounting firm should be especially cautious about entering sensitive client information into AI platforms without understanding the vendor’s data handling, retention, security, and training practices.

Valid Concerns Around AI-Generated Tax Advice

There are legitimate concerns around AI-generated written tax advice and client-facing recommendations.

AI systems can misinterpret nuanced tax scenarios, produce outdated or incomplete information, omit critical context, hallucinate legal or regulatory authority, or overstate confidence in uncertain conclusions.

For this reason, firms should be cautious about relying on AI-generated advice without professional review.

A responsible implementation approach may include restricting AI usage for high-risk activities, requiring review by licensed professionals, establishing internal AI usage policies, training staff on acceptable use, maintaining review procedures, and limiting exposure of confidential client information.

These are governance and risk-management issues, not evidence that AI is categorically incompatible with professional standards.

The Industry Needs More Nuance

The accounting profession should absolutely approach AI carefully and thoughtfully. Blind reliance on AI systems would be irresponsible. At the same time, declaring that AI tools are universally prohibited under Circular 230 is an overstatement that does not align with the actual language or structure of the regulation.

The more accurate position is this:

AI can be used responsibly within tax and accounting practices when accompanied by appropriate supervision, professional judgment, verification procedures, confidentiality safeguards, competence, diligence, strong internal controls, and AI governance.

AI should not be treated as magic. It should not be treated as a substitute for professional judgment. But it also should not be dismissed through broad claims that are not grounded in the actual regulatory text.

The future of AI in accounting will likely not be defined by prohibition. It will be defined by how effectively firms integrate these technologies into responsible professional workflows while maintaining the standards clients and regulators expect.

Responsible AI adoption starts with clear governance and strong client data protection. If your tax or accounting firm is updating its WISP, AI usage policy, or email security controls, Lappu AI can help you turn those requirements into practical implementation.

References