Malta’s tax authority says an AI-powered monitoring system helped it collect more than €1.6 billion in tax arrears since 2023, including €455.4 million by September 2026. The system flags possible compliance risks for auditors to review, giving a concrete example of AI being used inside public administration rather than merely promised on a strategy slide.
Watch Desk analysis
What happened
The Malta Tax and Customs Administration introduced the system, identified in MaltaToday’s account as SAS, in 2023. Commissioner for Tax and Customs Joe Caruana told the publication that auditors had entered more than 100 risk rules. One example checks whether a business’s declared sales for income tax match its VAT declarations; significant gaps can flag a case for follow-up.
The system also identifies late or incomplete filings and groups taxpayers by compliance behaviour. MaltaToday reports arrears collections of €374.8 million in 2023, €431.3 million in 2024, €419.7 million in 2025 and €455.4 million through the third quarter of 2026. The authority credits better case selection and improved compliance behaviour, while Caruana said human intervention and expert judgement remain central to decisions.
Why it matters
This is a sizeable, specific public-sector use of AI: data rules help decide which tax cases auditors examine, while officials retain a role in what happens next. The reported totals make the system’s stated reach tangible, but they do not by themselves show how much revenue the software caused to be recovered or how accurate its flags are.
The details matter beyond Malta. When an algorithm helps direct enforcement, the practical questions are what data and rules shape those decisions, how people can challenge mistakes, and whether human review is meaningful rather than ceremonial.
Our read
The useful story here is not that a computer has discovered a secret drawer labelled “tax money”. It is that Malta says automated risk checks are helping its tax authority select cases, with a commissioner describing both the rules and the continuing role of auditors. That is enough to make the deployment worth discussing, while leaving room to ask for evidence about accuracy, oversight and outcomes.
What to watch
- Whether Malta publishes more detail on the system’s risk rules and how often flagged cases prove relevant.
- How taxpayers can correct errors or challenge decisions informed by the system.
- Whether the authority reports future results alongside measures of accuracy, compliance and human review.
Discussion spark: When an AI risk system helps decide who gets audited, should the public be able to inspect the rules, or is that transparency incompatible with effective tax enforcement?
Sources and evidence
- AI helped taxman collect €1.6 billion in arrears (15 November 2026, 00:00 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.