AI Chatbots' Political Advice Questioned Ahead of Hungarian Elections
A recent study conducted in Hungary has raised concerns about the reliability of political recommendations provided by AI chatbots like ChatGPT and Gemini. The research, carried out in anticipation of the Hungarian parliamentary elections, found that these AI tools can generate inaccurate and inconsistent advice. Notably, the chatbots frequently provided different answers to the same questions when posed multiple times. Furthermore, they even suggested voting for political parties that were not actually listed on the official ballot.
As artificial intelligence tools become more integrated into daily life, their potential for influencing public opinion, particularly during electoral periods, warrants careful consideration. This study highlights the current limitations of large language models in providing accurate and consistent political guidance. The inconsistency in responses and the inclusion of non-existent options suggest that these systems may not yet possess the nuanced understanding or up-to-date, verified data required for such sensitive applications. Users should exercise critical judgment and verify information from AI sources, especially when making decisions with significant civic implications, recognizing that the underlying algorithms are trained on vast datasets that may contain biases or outdated information. The development of AI for civic engagement must prioritize accuracy, transparency, and robustness against misinformation.
AI-generated to prompt reflection — not editorial opinion, not advice, not a statement of fact. How this works.