Monday February 20, 2023 – This space has been pointing out that the Internal Revenue Service (IRS) is planning to devote its increased budget to machine learning and data analytics, rather than hiring thousands of new revenue agents.
This shift towards the use of artificial intelligence (AI) in making IRS decisions less transparent, as AI algorithms can be difficult or impossible to understand and their decisions are often cloaked in bureaucratic secrecy.
To protect American taxpayers Congress must insist on IRS algorithmic accountability. This includes publicizing pre-existing data sets used to train the algorithm, making available a detailed explanation of the algorithm’s rules and how it makes decisions, and providing a mechanism for taxpayers to challenge algorithms that appear to be arbitrary or unfair. Further, any AI decision made by the IRS should include some form of human review and oversight.
The use of artificial intelligence in government is unavoidable. However, it is essential that taxpayer rights remain paramount and that any AI decision has sufficient safeguards to protect them from being harmed by rogue algorithms. By increasing transparency and oversight of the IRS’s use of AI, Congress can ensure that taxpayers are not unfairly targeted or harmed by misguided algorithms.
Attorney Steven A. Leahy looks at AI algorithms and IRS decision-making on Today’s Tax Talk.