Automated Tax Planning: Who’s Liable When AI Gets It Wrong?

Abstract

Considering the continued proliferation of and rapid advancement in artificial intelligence technology, tax professionals are increasingly finding themselves confronted with novel accountability questions. If I render erroneous tax advice based on the output of an AI, to what extent will I be held professionally responsible? How do I navigate situations in which the AI’s tax analysis differs from my own, even if I struggle to document or even explain why I expect a different outcome? As AI becomes more powerful and is responsible for informing a greater number of important decisions, the challenge of assigning and apportioning this liability becomes progressively more difficult.

Inevitably, tax professionals will increasingly turn to AI-driven tools for assistance. Forty percent of legal professionals use or plan to use generative AI, and 50 percent believe that generative AI will transform legal practice. What do tax practitioners need to know before they dive into using generative AI in their daily work? We explore the novel liability considerations that will arise as professional firms implement AI tools for automated assistance with tax planning and analysis. It examines the ways in which AI frustrates and upends traditional legal concepts of liability by complicating who is liable when an AI informed analysis is flawed. It also explores approaches that regulators and tax practitioners can take to shield themselves against adverse AI consequences as they integrate computational tools into their workflow. Ultimately, it notes that the blame game is perhaps the same as it ever was—the responsibility for competent advice lies with the tax professionals who employ these and other tools.

Keywords: professional responsibility, tax, legal singularity, generative AI, professional liability

JEL Classification: H0