Navigating the Algorithmic Frontier: The Birth of the Council on AI Risk in Tax (CART)

The rapid integration of artificial intelligence into the global financial infrastructure has reached the tax sector, prompting a major collaborative effort to ensure that technological efficiency does not come at the cost of systemic integrity. The American Institute of CPAs (AICPA), in partnership with former IRS Commissioner Danny Werfel, has officially launched the Council on AI Risk in Tax (CART). This initiative serves as a centralized think tank designed to establish the guardrails for AI adoption in tax administration, compliance, and advisory services.

As AI transitions from a speculative curiosity to a strategic necessity, the tax ecosystem—comprising accountants, legal experts, technology developers, government regulators, and academics—faces a critical juncture. The formation of CART signals a proactive attempt to reconcile the immense efficiency gains of machine learning with the non-negotiable requirements of accuracy, ethics, and accountability.

The Genesis and Mission of CART

The Council on AI Risk in Tax is not merely a forum for discussion; it is a response to the accelerating pace at which AI is reshaping the tax landscape. The initiative brings together diverse stakeholders, including the National Association of Enrolled Agents and the Federation of Tax Administrators, to address the practicalities of implementation.

The primary objective of CART is to refine the "AI Risk Framework" published by the AICPA earlier this year. This framework acts as a foundational document for practitioners looking to integrate AI while mitigating potential liabilities. By convening cross-sector leaders, the council aims to create a robust repository of tools, guidance, and best practices that can withstand the scrutiny of both the IRS and the public.

Defining the Strategic Necessity

For many firms, AI is no longer an optional upgrade; it is a competitive requirement. However, the complexity of tax law, combined with the "black box" nature of some generative AI models, creates a unique risk profile. According to Melanie Lauridsen, the AICPA’s vice president of Tax Policy & Advocacy, the council is focused on "evolving the recently published risk framework" to ensure that as tax professionals adopt AI, they do so in a way that preserves public trust and institutional accountability.

Chronology of the AI Shift in Tax Practice

To understand the urgency behind CART, one must look at the rapid timeline of AI’s integration into the profession:

• 2023: The Proliferation Phase. Generative AI models began appearing in mainstream tax software, assisting with document summarization and research.
• Early 2024: Initial Caution. Professional bodies began warning about "hallucinations" in AI-generated tax research and the potential for privacy breaches.
• August 2024: Regulatory Stirrings. The IRS Office of Professional Responsibility issued preliminary guidelines concerning the use of AI, specifically touching upon the economic implications for taxpayers.
• September 2024: Calls for Clarity. The AICPA formally requested that the IRS clarify its stance on fee structures related to AI-driven efficiency, noting that existing language appeared to penalize firms for adopting technology.
• Present Day: The Formation of CART. Recognizing the need for a unified voice, the AICPA and Danny Werfel spearhead a formal coalition to bridge the gap between innovation and regulatory compliance.

Supporting Data: Why Risk Management is Paramount

The tax industry operates under a unique pressure: the stakes are high, and the margin for error is effectively zero. Unlike creative industries where AI errors might be perceived as stylistic quirks, a tax error can lead to severe legal and financial repercussions for clients.

The Integrity Gap

The primary risks identified by CART include:

1. Data Integrity: The risk that AI models may pull from outdated or incorrect tax code databases, leading to erroneous advice.
2. Privacy and Security: Protecting sensitive client financial data when inputting information into third-party AI platforms.
3. Accountability: The legal dilemma of who is liable when an AI model provides incorrect guidance—the software provider, the developer, or the licensed CPA?
4. Bias and Transparency: Ensuring that AI models used for tax administration or auditing do not unfairly target specific demographics or entities based on skewed training data.

The Council on AI Risk in Tax aims to address these issues through a series of scheduled meetings throughout the year, focusing on identifying "material risks" and developing a universal standard for risk mitigation.

Official Responses and Perspectives

The formation of CART has been met with significant interest from both the private sector and regulatory circles. Danny Werfel, whose leadership of the initiative adds significant weight to the council’s findings, emphasized that the industry cannot rely on self-policing mechanisms.

The Werfel Perspective

"Tax is high stakes, and we cannot assume that emerging AI tools will police themselves," Werfel noted during the announcement. His involvement signals a commitment to creating a bridge between the private sector’s technical capabilities and the public sector’s regulatory requirements. Werfel’s focus remains on refining the framework to ensure that practitioners can leverage the benefits of AI without exposing their clients to the pitfalls of unvetted, algorithmic decision-making.

The AICPA’s Stance

Melanie Lauridsen has been instrumental in framing the council’s work as a matter of "trusted guidance." The AICPA’s position is that without a unified, industry-wide governance strategy, the tax profession risks a fragmented adoption of technology that could lead to widespread errors and a decline in public confidence. By bringing together the "tax ecosystem," the AICPA hopes to foster a culture of responsible AI rather than a reactionary one.

Implications: The Battle Over Fee Structures

A significant point of contention currently driving the conversation around AI in tax is the recent guidance from the IRS Office of Professional Responsibility (OPR). The OPR suggested that tax practitioners should pass on AI-related cost savings—such as those gained from faster document processing or automated data entry—to their clients.

The Value-Pricing Conflict

The AICPA has pushed back against this suggestion, seeking formal clarification from the IRS. The institute argues that the OPR’s language ignores two critical realities:

1. The "Hidden" Costs of AI: Implementing AI is not free. It involves high subscription costs, investments in cybersecurity to protect client data, and the cost of training staff to use these tools effectively.
2. Value-Pricing Models: Many firms operate on a value-pricing model rather than a time-based model. If a firm uses AI to solve a complex tax issue in minutes that would have previously taken hours, the value provided to the client—the accuracy and the strategic outcome—remains the same or increases. The AICPA contends that charging based on "time saved" via automation is an outdated concept that overlooks the investment required to reach that level of efficiency.

This dispute highlights the broader tension between traditional regulatory interpretations of Treasury Circular 230 and the modern reality of digital transformation. Circular 230 prohibits "unconscionable fees," and the IRS is clearly concerned that practitioners might inflate fees while using cheap, automated tools. The AICPA’s goal via CART is to create a set of standards that makes the value of AI-enhanced services transparent, ensuring that fee structures are viewed as fair by both regulators and clients.

Looking Ahead: The Future of AI in Tax Administration

The work of the Council on AI Risk in Tax is just beginning. As the council meets throughout the year, its output will likely serve as the "gold standard" for firms navigating the digital transition.

Developing Tools and Insights

Beyond the risk framework, the council is tasked with:

• Developing Educational Modules: Training CPAs on how to audit AI outputs.
• Advocacy: Working with the IRS to refine guidance on fees and ethical AI use.
• Technology Assessment: Providing a vetted list of best practices for selecting and implementing AI software.

The implications of these efforts extend far beyond the accounting firm. They influence the taxpayer experience, the speed and accuracy of tax administration, and the overall stability of the financial system. As AI continues to evolve, the tax profession is positioning itself to be a leader in the responsible deployment of this technology.

Resources for Practitioners

The AICPA continues to offer support for firms currently navigating these waters. Their website provides introductory guidelines and a comprehensive FAQ section for those looking to implement AI in federal tax practice. For professionals, the message is clear: the integration of AI is inevitable, but the manner of its integration is a choice. Through the work of CART, the tax community is choosing a path of vigilance, collaboration, and professional rigor.

For those interested in contributing to the ongoing discussion or submitting ideas for future analysis, the AICPA welcomes engagement from all corners of the industry. The future of tax is being written today, and the Council on AI Risk in Tax is holding the pen.