Multiple studies show AI chatbots frequently give election-related information that is incomplete, outdated, or inaccurate, with about one-third of replies affected and notable gaps in Spanish-language responses; newer models vary in accuracy, with GPT-5.5 and Gemini 3.5 Flash among higher performers in some tests. Experts urge improved model reliability and up-to-date official resources, warning that overconfident or outdated AI outputs can erode voter trust and access.