Combating fraud and bias this tax season
COMMENTARY | By using cutting-edge AI and offering a modern self-service experience, state governments will be well on their way to weeding out fraud, improving compliance and closing the tax gap.
Taxes are the bread and butter of government, as they centralize the funds required for public projects and services. An effective tax system is one in which everyone pays what they owe—something that is, unfortunately, not often the case. The difference between taxes that are owed and taxes that are collected is called the tax gap.
To reduce the tax gap, government agencies should invest in technology that can effectively and affordably weed out tax fraud and non-compliance. With tax season in full swing, let’s dive into three strategies that state departments of revenue can use to minimize the tax gap.
1. Employ advanced analytics to flag probable fraud
As artificial intelligence continues to improve, so will governments’ ability to detect fraud and discrepancies in tax filings. In Arizona, for example, government leaders recently completed a pilot program to verify W-2 tax withholdings. In simplest terms, AI offers the ability to rapidly sift through mass amounts of data to pinpoint anomalies. By comparing historical data to existing filings, the state uncovered a $5 million discrepancy. But AI is only as smart as the data on which it is trained. Agencies are swimming in data—more than enough to build effective non-compliance and fraud detection models. The key is identifying the right methods and partners to put that data to use. With the right data and advanced AI techniques such as neural networks and beta-skeleton graphs, governments can even identify fraudulent schemes that cannot be detected by humans. These models also boast fewer false positives than other approaches.
2. Avoid bias and inequity in AI-enabled fraud detection
While AI is the technological cornerstone of closing the tax gap, agencies must beware of the possibility that bias will seep into their models. This can happen in two main ways: through proxy data (i.e., ZIP code data resulting in racially disparate impacts due to residential segregation) and historical data (i.e., training on a dataset that already has bias). Bias is particularly difficult to avoid when agencies rely on black box algorithms, which is why visibility is so important. ChatGPT is an example of AI that is precise but lacking in visibility. Instead, agencies should opt for models that use an open-source game-theory methodology, like the Shapely value, that clearly explains how much each data point contributes to the algorithm’s output. If a model flags an individual for tax fraud, the agency should be able to identify the contributing variables down to the percentage point.
3. Don’t overlook user experience
Sometimes, citizens fail to pay their taxes because the workflow or technology available to do so is not intuitive. In addition to investing in AI to flag anomalies and collect outstanding taxes, agencies should also work on improving the digital experience for citizens. By investing in a modern technology stack, taxpayers will have a user experience more like the ones they are used to from tech giants like Amazon and Netflix. From a single portal, taxpayers should be able to log in and do everything from filing returns and uploading documents to checking the status of a refund or audit. In short, a seamless self-service experience will help improve self-cure rates—meaning that citizens can address non-compliance independently—increase compliance and reduce the tax gap.
Ensuring that every public dollar gets collected is crucial to keeping government running smoothly at every level. By using cutting-edge AI and offering a modern self-service experience, state governments will be well on their way to weeding out fraud, improving compliance and closing the tax gap.
This 2023 tax season, it’s time for government decision-makers to assess the technology they currently have and to look ahead at savvy investments that could help reduce the tax gap even more. The time is now to crack down on tax fraud—and AI and user-experience are here to help.
Peter Arena is chief strategy officer at GCOM.