1. Home
  2.  | 
  3. Tax Audits
  4.  | Are AI tax audits the wave of the future?

Are AI tax audits the wave of the future?

On Behalf of | Apr 29, 2022 | Tax Audits

As engineers and programmers continue to refine technology, more and more organizations find new ways to use it to benefit their bottom line. Whether it is an automated assembly line featuring nothing but robots or machine learning programs geared toward uncovering potential compliance issues, numerous organizations find uses for advanced technology.

In these scenarios, the IRS is no different.

Many people worry that the use of artificial intelligence (AI) can lead to problems. They argue that the human element is what generally prevents cataclysmic errors during several types of procedures. The IRS has turned to AI for a combination of reaons. They face an enormous gap between the taxes that are paid annually versus the taxes they calculate to be owed. In recent years, that gap has ballooned to nearly half a trillion dollars – $400 billion. When combined with employee shortages, the IRS has turned to active-learning artificial intelligence to identify tax returns with red flags that suggest an audit could be necessary.

Is this an effective system?

In general, this system has been in place for decades. The IRS will feed a representative sample of returns into the computer and the program will analyze them line-by-line to uncover potential discrepancies. The government agency will then take those discrepancies and decide on next steps.

Unfortunately, the system hasn’t historically been proven all that efficient. The data yielded is of limited value and the IRS ultimately still relies on human agents to verify the machine data and decide if a structured audit is necessary. The solution that one research team suggested involves heavy reliance on active learning rather than simple programed parameters.

The team made up of a combination of Stanford researchers and IRS consultants have started developing a system where machine learning trains the IRS system to recognize issues and approach challenges in new ways. The active learning system will continuously update its parameters and learn based on various inputs. The system should identify red flags quicker with more efficiency. This model will likely require less human verification.

What does this mean for taxpayers?

In general, the average taxpayer should notice no difference. The AI system might help the IRS identify audit-necessary returns more quickly and efficiently than they could have in the past. A taxpayer who diligently files their taxes without errors, mistakes or exaggeration will likely avoid the random IRS audit. If an audit becomes a reality, however, it is wise to seek legal guidance to ensure you avoid costly penalties and devastating consequences.