What's Happening?
Danny Werfel, a former IRS Commissioner, has developed a comprehensive risk framework for the use of Artificial Intelligence (AI) in tax administration and preparation. This framework addresses the potential benefits and significant risks associated with
AI deployment in the tax sector. Werfel highlights that while AI offers substantial improvements in productivity and customer satisfaction, it also introduces complex risks that are not yet fully understood. The framework provides two distinct risk registers: one for tax authorities and another for tax preparers. These registers categorize risks into areas such as information integrity, fairness and legitimacy, security and data, institutional capacity for authorities, and technical, practice, legal and regulatory, business, and workforce/strategic for preparers. The initiative aims to guide organizations in responsible AI adoption by offering a structured approach to identify, assess, and mitigate these emerging risks, fostering a common understanding across the tax community.
Why It's Important?
The implementation of AI in tax administration and preparation carries significant implications for the U.S. tax system, affecting both government agencies and private practitioners. For tax authorities, the framework is crucial for maintaining public trust and ensuring equitable treatment of taxpayers. Risks like 'hallucination' (AI generating factually incorrect output) or 'bias in audit targeting' could lead to legal challenges, erode public confidence, and undermine the legitimacy of tax enforcement. For tax preparers, the framework addresses potential malpractice exposure, client data leakage, and reputational harm, which could severely impact their businesses and client relationships. The framework's emphasis on responsible innovation, starting with incremental deployments and thorough testing, is vital to prevent widespread errors and maintain the integrity of financial data. Without such a framework, the rapid adoption of AI could lead to unforeseen legal liabilities, financial inaccuracies, and a decline in public and professional trust in the tax system.
What's Next?
The proposed AI risk framework is intended to serve as a foundational tool for the tax community to collaboratively develop best practices and continuously improve as AI technology evolves. Tax authorities and preparers are encouraged to use these registers to conduct self-assessments of their readiness for responsible AI adoption and to perform due diligence on AI vendors. The framework suggests an eight-step methodology for deployment, starting with defining the scope of AI products, identifying applicable risks, prioritizing them, and assessing organizational governance and vendor accountability. Future steps will likely involve ongoing discussions among stakeholders, including the IRS, tax professionals, and AI developers, to refine the framework and establish industry-wide standards for AI use in tax. This collaborative effort aims to ensure that AI's transformative potential is harnessed while effectively managing its inherent risks.
Beyond the Headlines
The introduction of an AI risk framework in the tax sector touches upon deeper ethical and societal implications. The risk of 'policy drift,' where AI subtly reshapes tax rule interpretations without formal changes, raises concerns about democratic oversight and the potential for AI to inadvertently influence public policy. Furthermore, the framework highlights the risk of 'workforce and expertise erosion,' suggesting that over-reliance on AI could diminish human analytical capabilities and institutional knowledge within tax agencies and firms. This could lead to a future where complex tax matters are less understood by human experts, making the system vulnerable to sophisticated AI-driven errors or manipulations. The call for collaboration between tax authorities and preparers on shared risks underscores the need for a unified approach to address the ethical challenges of AI, ensuring that technological advancements serve the public interest without compromising fairness, transparency, or human accountability.
