How IRS and AI Are in a Love-Hate Relationship by Dr. Timothy Smith

Photo Source: Unsplash

AI has improved the capabilities of both the Internal Revenue Service (IRS) to identify tax fraud and for tax fraudsters to cheat the IRS and America's citizens. The IRS of the US Government processes millions of tax returns every year to bring in the over $3.3 trillion owed by private citizens and businesses combined. A massive problem for the IRS comes from a portion of the population that does not pay or underpays their taxes. According to the Government Accountability Office (GAO), some individuals and companies cheat on their taxes each year. The GAO reports that 85% of Americans pay their taxes in full and on time, garnering the IRS $2.8 trillion yearly. However, those cheating the system leave a $496 billion shortfall. Through investigation, litigation, and fines, the IRS recovers about $60 billion each year, but that still leaves the government short $428 billion, also known as the tax gap. (gao.gov)

To close the tax gap, the IRS has embraced AI to help speed up processes and automate the identification of tax cheats. AI provides the remarkable capability to examine millions of tax returns and financial records in a fraction of the time it would take a team of human auditors. AI truly shines in identifying patterns, and the IRS uses AI to identify fraud patterns such as underpaying taxes or taking unwarranted deductions. AI also helps to determine the best persons and companies for auditing. With reinforcement learning, AI will learn from its successes and mistakes to become more adept at identifying fraud.

For example, tax fraud by thieves filing under stolen identities robs people of millions of dollars by stealing their tax returns. Criminals have embraced AI to mount more realistic and sophisticated scams to steal from the IRS. The process involves the theft of PII or personally identifiable information such as date of birth, social security number, place of birth, and previous residence. With such information, a criminal can obtain a realistic-looking driver's license with the criminal's photo and the victim's address. With a driver's license, a criminal can create an IRS account that allows her access to previous tax filings and a way to file taxes with the IRS electronically. Many people have not opened an IRS account, leaving them vulnerable to a fraudster opening up one first in the victim's name. With the help of AI, the criminal can create a very clean tax filing that will not trigger an audit. The tax refund will then be transferred to the criminal's bank account. The Department of the Treasury uses AI to identify fraudulent checks and protect taxpayers. Last year, they claimed to have prevented $2.5 billion in high-risk transaction fraud and $1.0 billion in Treasury check fraud with the help of AI. (home.treasury.gov)

The IRS has a frenemy relationship with AI because it uses AI to prevent tax fraud better and close the tax gap that costs the government nearly half a trillion dollars annually. AI helps detect fraud and foil tax fraud scams. However, AI poses a growing threat to the IRS, with large language models and deep fake technology helping criminals create realistic scams that defraud taxpayers of their refunds. Practically, everyone should set up their IRS account, making it less vulnerable to being poached by a criminal. The IRS will continue to develop AI to manage the US tax collection better, and they will also have to manage the use of AI by bad actors with criminal intent.

Dr. Smith’s career in scientific and information research spans the areas of bioinformatics, artificial intelligence, toxicology, and chemistry. He has published a number of peer-reviewed scientific papers. He has worked over the past seventeen years developing advanced analytics, machine learning, and knowledge management tools to enable research and support high-level decision making. Tim completed his Ph.D. in Toxicology at Cornell University and a Bachelor of Science in chemistry from the University of Washington.

You can buy his book on Amazon in paperback and in kindle format here.