It has only been three years since OpenAI released the first publicly available ChatGPT. In that time, AI has changed the way professionals conduct their work. The tax profession is no stranger to this renaissance, as those working in tax now use this technology in many facets of their daily lives to promote more efficient and accurate work. It becomes more important now than even to understand how AI is being taken up by businesses, individuals, and the taxing authority. A recent academic study does just that by examining how AI investments enabled tax managers to integrate tax planning with business strategies in a way that yielded better corporate tax outcomes.
Taxes & AI
AI is already being used extensively by tax professionals. For instance, the IRS has taken up this technology to help provide more thoughtful scrutiny over whether and to what extent it audits taxpayers. They are not alone as state tax agencies have taken up AI to enhance their reach, especially given the fewer resources available to them and state revenue budget shortfalls.
What is less clear is how taxpayers have been using AI to enhance their ability to understand the tax code. Theoretically, corporations can use AI chatbots like ChatGPT and Gemini, to obtain a deeper understanding of the tax code and, by extension, beat the tax code by identifying tax planning opportunities. The demand for this type of AI has grown so substantially that we even now see the development and rise of specialized tax AI providers like Blue J, which is now being used extensively by tax professionals and accounting firms.
However, using AI in this way is easier said than done. Many caution taxpayers against relying too heavily on AI for tax purposes. AI is not always right, and, even if it was right based on the prompts provided, it can be wrong as a whole because it does not have all information available to make the correct decision. Lastly, privacy can be a concern when using AI. When a person puts prompts into ChatGPT or Gemini, that information is being put out into the world. If that information contains anything that is proprietary or confidential, there are few controls in place to prevent it from being spread, which could lead to confidential tax information becoming publicly available. Thus, there are questions as to whether and to what extent individuals and companies can benefit from a significant investment in AI.
Investments In AI Alter Corporate Tax Outcomes
Recently published research in the Review of Accounting Studies explores how firms’ early AI investments enabled managers to integrate tax planning with business strategies in a way that yielded better corporate tax outcomes. In a study titled “The use of artificial intelligence in decision-making: evidence from the effectiveness of corporate tax strategies”, researchers examined 9,786 firm-year observations across 1,433 unique firms and found that AI investments are associated with an increase in tax planning effectiveness. This article is co-authored by Trent Krupa of the Pennsylvania State University and Michelle Mullaney of Indiana University. Their study examines whether managers’ information processing constraints prevent the effective integration of tax strategies and business decisions — and whether AI mitigates those constraints.
In discussing the key motivation for examining their research question, Krupa states, “Today, firms generate unprecedented amounts of data in various forms, which are often dispersed across the firm. Despite the volume of available data, managers are human and subject to information processing constraints.” He continues to state, “Practitioners indicate that when executives make complex business decisions, there is a ‘data-insight gap. ‘”. This motivated us to better understand whether and how investment in AI, which is not subject to the same human limitations, is changing this data-insight gap.”
This study finds that firms’ early AI investments between 2010 and 2018 enabled managers to integrate tax planning with core business strategies better. Mullaney specifically notes that, “Corporate investment in AI tools, such as machine learning, is associated with more optimal tax decisions. This effect occurs through improvements to managers’ information and better capital management decisions.” The study specifically estimates that the stronger tax planning outcomes mean that one additional AI-skilled employee is associated with approximately $652K in additional after-tax earnings for the median firm.
The study also suggests that the findings are stronger among more operationally complex firms and when the firm’s tax managers are higher up in the firm. These additional findings combine to suggest that the results are more concentrated among firms that are more likely to benefit the most from improvements and when tax managers are more likely to be involved with core strategic business decisions.
While this study is among the first to consider how firms use AI to promote tax outcomes, the authors suggest there is more that can be done. Krupa specifically states, “First, the results suggest that improving the service operations of the IRS may have spillovers into business creation, which may not currently be considered in the costs and benefits of IRS operations. Second, they suggest that simplifying tax filing and compliance by reducing tax complexity may encourage entrepreneurship. Third, the evidence that Taxpayer Assistance Centers mainly affect local entrepreneurship suggests that improving awareness of Taxpayer Assistance Centers in neighboring geographic areas may be a low-budget alternative to expanding Taxpayer Assistance Centers operations.”
These findings have clear implications for firms that are currently considering integrating AI into their workplace. For starters, a significant investment in AI may be costly. These costs include those that are direct, such as investing in a subscription for employees to have access to Blue J, or indirect, such as the costs to train employees on how to effectively and efficiently use AI in the workplace. However, this study’s findings suggest that the investment may be positive net present value when compared to the benefits derived from enhanced tax efficiency. However, the study’s findings also caution that the positive impacts may be reliant on processes and investments already in place. Either way, AI is a technology that is here to stay, and firms may wish to embrace this technology before they fall behind their peers and competitors.
