AI Chatbots Dispute Election Lies But Illustrate Them

AI chatbots, including ChatGPT, Gemini, Grok, Claude, Perplexity, and DeepSeek, have demonstrated a capacity to dispute election conspiracy theories when directly confronted with them, according to a new study by the Brennan Center for Justice. Researchers from the Brennan Center posed five false claims central to recent attacks on election integrity to these chatbots between February and August of this year. These claims included assertions such as "Noncitizens vote in large numbers" and "Voting machines are rigged." Even when researchers applied repeated pushback and scrutiny, the AI models consistently refused to validate these false narratives. However, the study also revealed significant issues with the accuracy of the chatbots' responses. Approximately half of all responses contained an inaccuracy or a faulty citation, and one in three responses included a factual error, a broken web link, or a misleading citation. Despite these accuracy concerns, the chatbots were described by researchers as "behaved as sympathetic listeners." A more concerning finding emerged regarding the image and video generation capabilities of AI tools developed by the same companies behind these chatbots. The study found that ChatGPT, Gemini, Grok, and Meta's platforms were more amenable to generating images that supported or illustrated conspiracy theories and other forms of disinformation. For instance, researchers were able to prompt these tools to create images depicting scenarios such as an individual collecting mail ballots in a "ballot collection" bag, thereby illustrating the concept of "ballot harvesting." In some instances, the AI-generated images included "convincing details," further enhancing their potential to mislead viewers. This dual behavior highlights a critical challenge: while conversational AI may be programmed to reject misinformation, the generative AI tools from the same providers can be used to create visual content that actively promotes it. The study underscores the need for robust safeguards and ethical considerations in the development and deployment of AI technologies, particularly those capable of generating persuasive visual content that could be exploited to spread disinformation and undermine democratic processes. The Brennan Center for Justice, a nonpartisan law and policy institute, focuses on strengthening democracy and the rule of law, and its research aims to inform policy decisions and public understanding of critical issues.