Given the creeping encroachment of AI into almost every aspect of modern life — from making music to helping Google maintain even more of a monopolistic chokehold on the open web — it was probably to be expected that the federal government would eventually decide to employ the technology in service of tax collection.
IRS commissioner Danny Werfel, aka the top federal official in charge of squeezing as much money as possible from taxpayers, said in public remarks just a few days ago in Washington DC that the tax agency is exploring two use cases for AI: Deploying the technology in the form of chatbots, to answer questions from taxpayers. And also to help the agency identify people who aren’t paying Uncle Sam enough money.
“Right now, I believe that there are AI solutions that we have not yet leveraged that exist today that can help with some of these basic questions to the benefit of taxpayers,” Werfel said. “And on the other side of the equation, we are using AI today to do even more to unlock and spot this complexity.”
Werfel’s remarks came days after Tax Day 2024 (April 15) and also against the larger backdrop of generative AI becoming increasingly mainstream thanks to companies like OpenAI, Meta, and Google baking the technology into their core products. All that said, however, I can’t help but feel like there’s something dystopian about the government turning to AI for a core function like this. Were this to be deployed at scale, I would expect at a minimum there to be a corresponding reduction in the IRS’ headcount.
You’re there to do a job, in other words. And if AI is now doing more of it for you, there ought to be a cost savings that flows from that (even though, who am I kidding, this is the federal government we’re talking about — the place where efficiency and logic go to die). “We have to be cautious and thoughtful with how we deploy AI,” Werfel continued.
“We have a fundamental responsibility that is primary for us, and that is to protect taxpayer rights. And those rights include things like the right to privacy and the right that we will not add unnecessary intrusion into that privacy. We need to stay 1,000 miles away from even the perception that AI is in any way violating the responsibility we have to protect taxpayer privacy.”
