Artificial intelligence can be a valuable tool for handling complex tax issues. However, those who rely on the answers of a general AI system and consequently make a tax error cannot automatically expect leniency. A recent ruling by the Federal Fiscal Court demonstrates that a legal error caused by AI is generally attributed to the taxpayer.
Taxpayers are increasingly turning to freely available AI systems like ChatGPT or similar applications to solve complex tax problems. The problem is that general AI systems are not necessarily specialized in Austrian tax law. They can therefore, for example, draw on foreign legal sources, apply outdated legal principles, or cite non-existent court decisions and legal literature. In the specific case at hand, a researcher had prepared her legal documents using AI. The procedural approach suggested by the AI later proved inadmissible. In the proceedings before the Federal Fiscal Court (BFG), the appellant argued that she had been subject to a legal error caused by the AI. However, the BFG rejected this argument. The legal error was attributed to the taxpayer and therefore disregarded. She was also denied a retrial (BFG Vienna, RV/7101070/2026).
The ruling clarifies that the use of AI does not, in principle, eliminate the taxpayer's personal responsibility. Anyone who relies independently on AI-generated information regarding tax matters cannot claim excusable error of law simply because the technology provided incorrect information. The situation may be different when advice is sought from a qualified professional. According to the Supreme Court's jurisprudence, a taxpayer enjoys special protection if they fully describe the relevant facts to a qualified advisor and receive advice that is not obviously incorrect. This can be particularly important in cases of tax liability, reinstatement of prior status, or even in tax criminal law. However, even lawyers and tax advisors cannot simply accept AI-generated content without scrutiny. The Supreme Court has already clarified that qualified professionals must review AI-generated documents for their accuracy. Erroneous or "hallucinated" legal sources can therefore have liability and professional consequences for the advisor as well.
AI can significantly simplify tax preparation, but it does not replace expert review or the taxpayer's personal responsibility. Anyone who relies directly on an erroneous AI-generated answer generally bears the risk of its errors. AI can support tax advice – but it is no substitute for professional tax or legal counsel.
BFG June 10, 2026, RV/7101070/2026