Germany is considering expanding the use of artificial intelligence across its tax administration system as part of proposed legal reforms designed to modernize tax enforcement and improve fraud detection. The plans remain under review and have not yet been formally adopted.
According to reports, Germany’s Finance Ministry is preparing changes through the proposed Annual Tax Act 2026 that would allow tax authorities to use larger volumes of real taxpayer data to train AI systems. The proposal would also relax some data retention rules, enabling certain taxpayer information to remain available for AI development for up to one year before deletion.
German tax authorities already use various forms of automation and algorithmic analysis under existing legislation. Since the Taxation Modernisation Act took effect in 2016, automated systems have increasingly supported tax assessments, risk analysis, audit prioritization, and fraud detection. The new proposal would significantly broaden how AI systems are trained and integrated into these processes.
Under the planned framework, AI systems could analyze large datasets to identify unusual financial patterns, flag potentially fraudulent filings, and help authorities prioritize high-risk cases for human review. Supporters argue the technology could improve efficiency as tax agencies face rising workloads and increasingly complex financial data.
Researchers also note that German tax authorities already use automated tools to collect and compare publicly available information from online marketplaces, websites, and social media platforms against submitted tax records. AI-assisted analysis could potentially expand these capabilities further if the proposed reforms move forward.
However, the proposal has generated criticism from privacy advocates, legal experts, and digital rights groups. Critics warn that using real taxpayer information to train AI systems could create risks involving surveillance, profiling, data misuse, and algorithmic bias. Some experts also argue there is limited transparency around how certain automated tax systems currently operate and how risk scores are generated.
Privacy researchers have called for stronger oversight and clearer safeguards before any expanded AI deployment is approved. Concerns include how long taxpayer information would be retained, who would have access to training data, and whether individuals would be informed when automated systems influence audit or enforcement decisions.
The German proposal remains subject to political debate and legislative approval. Officials have not finalized the framework, and additional changes could still be introduced before any nationwide rollout occurs.