Navigating the Algorithmic Frontier: AICPA and Industry Leaders Launch Council on AI Risk in Tax (CART)
The rapid integration of artificial intelligence (AI) into the global financial infrastructure has reached a critical inflection point for the tax profession. As tax administrators and practitioners scramble to harness the efficiency of generative AI and machine learning, the foundational requirements of the profession—integrity, accountability, and accuracy—face unprecedented challenges.
In response to this shifting landscape, the American Institute of CPAs (AICPA), in collaboration with former IRS Commissioner Danny Werfel, has spearheaded the formation of the Council on AI Risk in Tax (CART). This cross-disciplinary initiative aims to bridge the gap between technological innovation and the rigorous governance required to maintain public trust in the tax system.
The Genesis of CART: A Strategic Necessity
The launch of CART represents more than a collaborative committee; it is a defensive and proactive maneuver designed to standardize the "rules of the road" for AI in tax. The council convenes a diverse assembly of stakeholders, including representatives from the National Association of Enrolled Agents, the Federation of Tax Administrators, legal scholars, technology experts, and academic researchers.
The core mission of CART is to refine the recently published AI risk framework, a document intended to serve as a baseline for responsible AI adoption. As Danny Werfel, a key architect of this initiative, noted, the stakes are simply too high to leave AI integration to the invisible hand of the market.
"Tax is high stakes, and we cannot assume that emerging AI tools will police themselves," Werfel remarked. "The AI risk framework published earlier this month provides a starting point. Through this new council, the AICPA is bringing the tax community together to refine that framework and explore additional tools that can help tax administrators and practitioners realize the benefits of AI while managing the various and material risks that AI presents."
Chronology: The Escalation of AI in the Tax Profession
The formation of CART follows a period of accelerated AI adoption that has outpaced formal regulatory guidance. Below is a timeline of the recent developments necessitating this oversight:
- 2023–2024 (The Adoption Phase): Tax firms and government agencies began integrating Large Language Models (LLMs) and automated data processing tools to handle high-volume compliance tasks and predictive audit modeling.
- Early 2024 (Identifying the Gap): Industry leaders recognized that while AI drastically reduced processing time, it introduced significant risks regarding data privacy, "hallucinations" in tax research, and potential biases in algorithmic audit selection.
- September 2024 (IRS Guidance): The IRS Office of Professional Responsibility issued guidelines under Treasury Circular 230, emphasizing that tax practitioners must ensure fees remain conscionable and that AI-related cost savings should be passed on to clients.
- Late September 2024 (Industry Pushback): The AICPA publicly sought clarification on the IRS guidelines, arguing that the current language ignores the substantial capital investment and risk-management costs inherent in implementing enterprise-grade AI.
- October 2024 (The Launch of CART): The AICPA officially announced the establishment of the Council on AI Risk in Tax to formalize best practices and provide a unified voice for the profession.
Supporting Data: Why Governance is Non-Negotiable
The urgency behind CART is backed by a shift in how practitioners view technology. According to recent AICPA internal assessments, AI is no longer categorized as a "competitive advantage" but as a "strategic necessity."
The Risk Matrix
The council is currently focusing on three primary pillars of risk:
- Information Integrity: Ensuring that AI tools, which often rely on probabilistic data, produce deterministic tax outcomes that satisfy the high standards of accuracy required by the IRS.
- Accountability: Defining who is responsible—the developer, the firm, or the individual CPA—when an AI-driven filing leads to an error or a breach of taxpayer confidentiality.
- Risk Management: Developing guardrails against "algorithmic drift," where AI models inadvertently change their logic over time, potentially leading to inconsistent tax treatments across different client demographics.
Melanie Lauridsen, the AICPA’s Vice President of Tax Policy & Advocacy, underscored that the council’s work is essential to prevent a crisis of confidence. "AI has the potential to transform how tax is administered and practiced but requires thoughtful governance, practical risk management, and trusted guidance," Lauridsen stated. "That is why this council was launched—to develop the tools and insights needed to adopt AI responsibly while preserving trust."
Official Responses and the Conflict over Value Pricing
A central point of friction currently occupying the tax community is the relationship between AI-enabled efficiency and fee structures. The IRS, citing Treasury Circular 230 (31 C.F.R. Part 10), has signaled that practitioners should not charge fees that are "unconscionable" and must account for the efficiency gains provided by AI.
The AICPA’s formal response to this directive has been firm. In their request for clarification, they have raised two critical points:
- The Myth of Simple Cost Savings: The AICPA argues that the IRS directive fails to acknowledge the high cost of subscription models, data security protocols, and the human oversight required to verify AI output.
- Value-Pricing vs. Cost-Plus: The profession has largely moved toward "value pricing"—charging for the expertise and outcome—rather than a "cost-plus" model based on hourly time. The IRS language, they contend, risks forcing the profession into an outdated pricing model that punishes firms for investing in innovation.
CART is expected to play a vital role in drafting a white paper that explains these nuances to regulators, effectively bridging the gap between tax authorities and the private sector.
Implications for the Future of Tax Practice
The implications of the Council on AI Risk in Tax extend far beyond immediate compliance. By standardizing risk management, CART is effectively setting the professional standards for the next decade of tax administration.
Implications for Small and Mid-Sized Firms
For smaller accounting firms, the risk of "AI-illiteracy" is high. These firms often lack the resources to build proprietary risk models. CART’s work will likely provide a "plug-and-play" framework, allowing smaller practices to leverage AI without exposing themselves to professional liability or IRS scrutiny.
Implications for Tax Administration
The IRS itself is in the midst of a massive technological modernization effort. As the agency uses AI to detect fraud and select audits, the feedback loop between the IRS and practitioners becomes crucial. CART provides a formal mechanism for practitioners to signal where AI-driven government audits may be overreaching or malfunctioning, fostering a more transparent administrative environment.
Implications for Professional Liability
Insurance carriers are already beginning to scrutinize how accounting firms use AI. A firm that adopts the guidelines established by CART will likely be in a much stronger position regarding professional liability insurance, as they will have documented evidence of following industry-recognized "best practices."
Looking Ahead: The Road to Standardization
As CART begins its cycle of meetings throughout the year, the tax profession will be watching closely. The council is tasked with moving beyond the abstract concepts of AI ethics and into the "nitty-gritty" of implementation. This includes creating checklists for firm owners, drafting model engagement letters that disclose the use of AI, and creating a unified taxonomy for AI risks in the tax space.
The AICPA has already provided introductory guidelines for responsible AI use, which serve as a foundational document for the council’s work. However, as the technology evolves, so too must the council. The goal is a dynamic, living framework that can adapt to new breakthroughs in Large Language Models and predictive analytics.
Conclusion: Trust as the Ultimate Currency
At its heart, the tax system is built on trust—the trust of the taxpayer that their data is safe, and the trust of the government that the revenue collected is accurate. By bringing together the brightest minds in accounting, law, and technology, CART is ensuring that as we transition into an era of machine-assisted tax practice, that trust remains intact.
For those interested in participating or tracking the progress of these initiatives, the AICPA maintains a repository of resources, FAQs, and updates regarding federal tax practice and AI. The formation of CART is a significant step toward ensuring that the digital transformation of the tax profession is not only efficient but also resilient, ethical, and fully aligned with the public interest.
For more information on the Council on AI Risk in Tax or to contribute to the ongoing dialogue, practitioners are encouraged to consult the official AICPA resources or reach out to the organization’s tax advocacy department. Readers interested in suggesting future topics for exploration are encouraged to contact Martha Waggoner at [email protected].
