OpenAI Cribbed Our Tax Example, But Can GPT-4 Really Do Tax?

The authors explain where OpenAI got the tax law example in its livestream demonstration of GPT-4, why GPT-4 got the wrong answer, and how it fails to reliably calculate taxes.

cs.AI cs.CL Law Economics Computer Science

View graph Source

Paper

Similar papers

Generative AI as Tax Attorneys: Exploring Legal Understanding Through ExperimentsTomasz Strak2024 · EUROPEAN RESEARCH STUDIES JOURNAL In Library

Is ChatGPT an Accurate Source of Information for Uninformed Taxpayers?Hannah Smith Antinozzi, Lauren Cooper2025 · Journal of Emerging Technologies in Accounting In Library

Shaping the Future of Tax Advisory Using Artificial IntelligenceConor J. Kelly2024 · International Journal of Science and Research (IJSR) In Library

ChatGPT-4: Can It Handle Real-World Accounting Cases?Theresa F. Henry, David A. Rosenthal, Rob R. Weitz2025 · Journal of Emerging Technologies in Accounting In Library

Large Language Models' Understanding of Math: Source Criticism and ExtrapolationRoozbeh Yousefzadeh, Xuenan Cao2023 · arXiv (Cornell University) · 6 citations In Library

Large Language Models' Understanding of Math: Source Criticism and ExtrapolationRoozbeh Yousefzadeh, Xuenan Cao2023 · Research Square In Library

GPT as Knowledge Worker: A Zero-Shot Evaluation of (AI)CPA CapabilitiesJillian Bommarito, Michael James Bommarito, Daniel Katz et al.2023 · arXiv (Cornell University) · 68 citations In Library

Performance of LLMs on VITA test: potential for AI-assisted tax returns for low income taxpayersSina Gogani-Khiabani, Ashutosh Trivedi, ShinPing Chyi et al.2025 · Artificial Intelligence and Law In Library

Unleashing efficiency and insights: Exploring the potential applications and challenges of ChatGPT in accountingJoanna Zhao, Xinruo Wang2023 · Journal of Corporate Accounting & Finance In Library

A First Look at Financial Data Analysis Using ChatGPT-4oWen-Hsiu Chou, Zifeng Feng, Bingxin Li et al.2025 · Journal of risk and financial management In Library

Source of record

arXiv

44B

Welcome.

This panel sticks with you. Pick anything below and it opens right beside it, so you can dig through 60,000-plus records without ever losing your spot here.

Travel 44B

LibraryPapers, policy, standards, statute LabsEvery organization building AI ModelsIntelligence, price, and speed BenchmarksThe evaluation catalog IncidentsAI failures and harms SearchOne field across everything DashboardThe state of AI accountability in NY 44B RegistryThe Art. 44-B compliance portal