AI at the IRS: Transforming Tax Enforcement and the Future of Taxpayer Services

While artificial intelligence (AI) has been around in some form since 1966, recent improvements have resulted in an explosion of its use. Many of us use it to automate tasks, thereby boosting efficiency while reducing the risk of human error. In fact, the title of this article was written by AI. The IRS has also made recent strides in utilizing AI. As of January 2026, the IRS has experienced a workforce reduction over the past year of approximately 25% and is facing further budget cuts to taxpayer services and enforcement in its Fiscal Year 2026 budget. To limit the impact of this on the IRS’ key functions, AI is now being employed in audits, fraud detection, and taxpayer services. 

AI in Audits 

The IRS uses AI in both the selection of tax returns for audit, as well as to conduct the audit itself. It is used to identify high risk returns, especially those filed by large corporations; complex partnerships; high wealth individuals, and users of digital assets. AI has been incorporated into the IRS’ Discriminant Function System (DIF), which uses an algorithm based on a large database of returns and a specific formula. The goal is to identify those returns with the highest potential for inaccuracies and discrepancies between income and deductions. DIF scores tax returns using various criteria and assigns each criterion a different numerical score. Returns with higher DIF scores are selected for closer review to determine if an audit is warranted. The use of AI in audit selection is aimed at improving efficiency and fairness by selecting returns based on objective standards. 

Once a return is selected for audit, the IRS also employs AI in the examination itself. Revenue agents are beginning to have access to generative AI programs to assist with routine drafting tasks such as information document requests and exam reports. These programs can analyze large volumes of information to create first drafts, which revenue agents review and finalize. AI is also being used in data integration to process and analyze large amounts of data in various formats to extract the most relevant information and present it in an easily understood format for revenue agents. The continued use of AI in examinations will allow revenue agents to streamline routine tasks so they can focus on more complex areas of the audit, potentially resulting in quicker audits.  

AI in Fraud Detection 

The IRS has also begun employing AI to detect fraud by using it to spot emerging compliance threats, with the goal of implementing real-time AI-based checks during the tax return filing process. To that end, in January 2024 the IRS issued a request for information seeking assistance to better understand how AI capabilities could be used to identify issues with tax returns at their initial filing. Based on the identification of these issues, the IRS would be able to contact taxpayers and allow them to quickly make corrections, avoiding the need to file an amended return or undergo a full audit. The IRS has also stated that they would use AI to notify taxpayers of potential credits or deductions to which they may be entitled but did not claim on the return. Finally, the use of AI at the time of filing would allow the IRS to more quickly identify common tax schemes and warn the public before they unknowingly become involved with an unscrupulous tax promoter.. Finally, the use of AI at the time of filing would allow the IRS to more quickly identify common tax schemes and warn the public before they unknowingly become involved with an unscrupulous tax promoter. 

AI in Taxpayer Services 

The IRS has repeatedly reinforced its commitment to modernizing tax administration. A key element to the modernization process is improving taxpayer service via communication and enhanced access to information. The IRS uses AI-driven voicebots and chatbots to answer common taxpayer questions, route calls to the appropriate group, and provide information in real time.  

Additionally, in late 2025, the IRS rolled out a new AI program, known as Agentforce, across the Office of Chief Counsel, Taxpayer Advocate Services, and the Office of Appeals. Agentforce is intended to search documents and provide case summaries, allowing the personnel in these groups more time to focus on personal interactions with taxpayers in resolving complex casework and taxpayer disputes. 

Concerns with the Expanded Use of AI 

As AI becomes more prevalent in society and the demands on a reduced IRS workforce intensify, the IRS’ expanded use of AI is inevitable. However, groups have voiced concerns regarding its use as AI continues to evolve. One agency watchdog, the Government Accountability Office (GAO), has issued multiple reports highlighting the potential for unintended bias by AI in selecting returns for exams. For instance, independent studies have confirmed that Black taxpayers are audited at a rate three to five times higher than others. The GAO identified “unintentional algorithmic biases” as a possible source for this disparity. As explored by the GAO and the Treasury Inspector General for Tax Administration (TIGTA), AI programs are created using pre-existing data. To the extent this data has been impacted by biases and social inequities, the resulting AI program may continue to perpetuate the disparities.  

As the IRS uses artificial intelligence to handle the challenges of a reduced workforce and budget cuts, taxpayers need to be aware of its impact on audits, fraud detection, and taxpayer services. Taxpayers contacted by the IRS should engage a trusted tax advisor to help them navigate the process.