Business
Consumers turning to artificial intelligence for financial guidance may be getting dangerously bad advice. A recent report has found that leading AI chatbots deliver incorrect responses to money-related queries far more often than they get them right. The study tested several widely used conversational AI tools on a range of personal finance and tax questions. In many cases, the chatbots not only supplied wrong answers but also invented rules that do not exist—a phenomenon known as hallucination. Perhaps more troubling, some systems failed to account for tax law changes that are already scheduled to take effect, meaning their advice could be outdated before users even act on it. Researchers behind the report say the findings highlight a fundamental flaw in relying on generative AI for high-stakes financial decisions. Unlike a search engine that points to authoritative sources, a chatbot generates plausible-sounding text that may be completely disconnected from reality. When the topic involves money, penalties, or deadlines, such errors can lead to real financial harm.
The report does not name every chatbot tested, but it notes that the problem appears across multiple major platforms. Even models that perform well on general knowledge tasks struggled with specific, rule-based financial questions. That gap suggests the issue is not simply a lack of data but a deeper limitation in how these systems reason about regulated domains. For now, experts recommend treating AI chatbots as a starting point for research rather than a final authority. Anyone with a serious financial question should verify answers with a qualified professional or an official government source. The convenience of instant AI answers, it seems, comes with a significant risk of being confidently wrong.