Purbaya to Utilise AI to Detect Potential Tax Shortfalls
Finance Minister Purbability Yudhi Sadewa has emphasised the intention to utilise Artificial Intelligence (AI) technology within the National Single Window (LNSW) to detect potential untapped state revenue, including indications of tax underpayments.
He explained that the system leverages the integration of various data available in the LNSW to assist analysts in conducting more in-depth inspections. This integration is considered vital as manual data processing is deemed difficult to sustain as information volumes continue to grow.
“We are developing an application here using AI as well. To assist our analysts in identifying potential revenue that could be obtained through system improvements,” Purbaya told reporters in Jakarta.
According to Purbaya, the system will be developed to eventually identify potential tax shortfalls at the corporate level. This data will be cross-referenced with taxpayers’ Annual Tax Returns (SPT) and can be traced back several years.
The government can subsequently request tax authorities to conduct audits if companies are found to be indicated as having failed to meet their tax obligations. Additionally, he will utilise this technology to pursue potential tax revenue from illegal gold trading.
“It will eventually reveal who is making unrecorded purchases or buying goods from illegal mining (PETI). We will pursue that. Revenue that has not yet entered the state coffers will be pursued,” he added.
The Finance Minister has not yet detailed the potential revenue that can be gathered through this process. However, he stated that he has already secured data on the top ten players in illegal gold trading that can be investigated.
“Once the names of the companies (PT) are identified, we will pursue them. I will instruct tax officials to audit those companies, where they sourced their supplies, and trace it back. If they have not paid tax, they must pay immediately. This is what is meant by extensification—increasing the efficiency of tax collection through data analysis,” he concluded.