ChatGPT, an artificial intelligence chatbot, is generating buzz in legal academia and practice. On November 30, 2022, OpenAI made ChatGPT available for public use through an open beta. In less than one week, the system was deployed by more than 1 million registered users. Driving this extremely rapid adoption were claims — and plenty of evidence — of impressive algorithmic feats performed by ChatGPT. The capabilities of this technology reportedly include achieving passing scores on practice bar and medical board examinations. Unsurprisingly, many law schools are discussing how to cope with students' using this technology to potentially cheat in their writing and exams. On the other hand, TaxBuzz said something the tax community would be glad to hear:ChatGPT's tax advice was wrong 100% of the time.
Given the divided prospect, the question arises: what are AI's current capabilities in tax analysis? Benjamin Alarie, a law professor at the University of Toronto and the CEO of Blue J Legal Inc., wrote an interesting essay, :The Rise of the Robotic Tax Analyst, 178 Tax Notes Fed. 57 (2023), discussing AI's role in tax research and analysis by 2030. As the title suggests, the content is indeed interesting. However, what I find more interesting is that Alarie wrote the essay with Generative Pre-Trained Transformer 3 (GPT-3). He assessed the role of GPT-3 as relatively light, limited primarily to sections III through V. Alarie also heavily edited the text produced by GPT-3, so he did not identify the text it contributed or declare it a co-author. Nevertheless, it was a shocking experience to read a sophisticated work product produced by ChatGPT. Readers looking at GPT-3’s contributions in sections III through V may notice that the content aligns with ChatGPT's work product in other media, such as collecting and processing existing information and using comparisons, inferences, and speculation to analyze and draw conclusions.
Alarie's essay is not the first attempt to write a legal essay with ChatGPT. Alarie and the late tax professor Arthur Cockfield of Queen's University wrote a peer-reviewed law review article; Will Machines Replace Us? Machine-Authored Texts and the Future of Scholarship, 3 L. Tech. & Hum. 5 (2021). The article records GPT-3 as a co-author and allows readers to see the machine's textual analysis uninterrupted and unedited by human beings. Cockfield and Alarie ultimately conclude that “although GPT-3 is not up to the task of replacing law review authors currently, we are far less confident that GPT-5 or GPT-100 might not be up to the task in future.” Additionally, within days of the release of ChatGPT, Andrew Perlman, the dean and a law professor at Suffolk University, prompted the program to write an essay on how ChatGPT is likely to influence the future of law. The abstract to the essay, which Perlman wrote himself, concludes that “the disruptions from AI’s rapid development are no longer in the distant future. They have arrived, and this document offers a small taste of what lies ahead.” Even beyond GPT-3 and its progeny, machine learning and AI are beginning to make their influence felt in unexpected places, and there is little reason to expect these and related developments to abate at the doorstep of tax analysis.
The remainder of the essay revisits Blue J's predictions on actual cases and outlines some new ways machine-learning technology might affect tax research and analysis. AI-driven legal research tools can generate detailed diagrams of complex transactions, highlighting connections and implications that individuals might overlook. Furthermore, natural language processing (NLP) and large language models will increasingly enable research platforms to quickly scan and extract key concepts from various sources, helping practitioners better understand their implications. The essay also predicts that governments will use AI to combat aggressive tax planning and detect and prosecute tax evaders. Lastly, tax practitioners will use AI to accurately predict the outcome of tax disputes and provide more competent tax advice. Whether you agree or not, the essay gives us an excellent opportunity to anticipate how AI could affect tax analysis and research and consider why it is vital to begin turning our attention to the possibilities.