# Can LLMs Identify Tax Abuse?

## TL;DR

This paper explores the capability of large language models to understand, verify, and generate complex U.S. tax-minimization strategies, revealing their potential to uncover novel approaches and assist in combating tax abuse.

## Contribution

It demonstrates that advanced LLMs can interpret, verify, and create detailed tax strategies, including discovering a new strategy, showcasing their potential in real-world legal domains.

## Key findings

- LLMs can interpret complex tax laws
- LLMs can fill gaps in strategies
- LLMs identified a novel tax strategy

## Abstract

We investigate whether large language models can discover and analyze U.S. tax-minimization strategies. This real-world domain challenges even seasoned human experts, and progress can reduce tax revenue lost from well-advised, wealthy taxpayers. We evaluate the most advanced LLMs on their ability to (1) interpret and verify tax strategies, (2) fill in gaps in partially specified strategies, and (3) generate complete, end-to-end strategies from scratch. This domain should be of particular interest to the LLM reasoning community: unlike synthetic challenge problems or scientific reasoning tasks, U.S. tax law involves navigating hundreds of thousands of pages of statutes, case law, and administrative guidance, all updated regularly. Notably, LLM-based reasoning identified an entirely novel tax strategy, highlighting these models' potential to revolutionize tax agencies' fight against tax abuse.