Analysis suggests that ChatGPT and Gemini may be subtly influencing your voting choices.

      Your AI chatbot also has a political bias

      It’s already quite common to seek assistance from AI chatbots for tasks like emails, homework, travel arrangements, and much more. Thus, it was only a matter of time before politics came into play. A recent analysis by The Washington Post indicates that popular AI chatbots may not be as politically neutral as they seem. The Post evaluated the algorithms behind OpenAI’s ChatGPT, Google’s Gemini, Anthropic’s Claude, DeepSeek, xAI’s Grok, and Gab’s Arya using a series of political questions aimed at assessing how these chatbots respond to contentious topics.

      The report found that OpenAI’s model provided predominantly left-leaning responses in 80% of the cases, while Google’s Gemini generally adopted a both-sides perspective, presenting left- and right-leaning arguments in over 90% of its responses.

      This is not evidence of voter manipulation

      It’s crucial to clarify that AI is not directly influencing election outcomes. The report does not demonstrate that chatbots are altering user voting behaviors. The analysis focused solely on brief answers to political inquiries and not on actual voting actions. However, these results are significant. When users inquire about campaign finance, healthcare, policing, immigration, taxes, or the Electoral College, the responses they receive can shape their understanding of the issues before they engage with news articles or hear from candidates. The way these issues are framed matters, as chatbots present their answers with a sense of confidence.

      The Post also observed that even AI models branded as conservative or anti-"woke" did not always perform accordingly. Grok did provide more right-leaning responses compared to other bots but still leaned toward left-leaning arguments on several occasions. Gab’s Arya, which promotes Christian and conservative values, similarly often leaned left in its responses during the evaluation.

      Companies assert that neutrality is the aim

      The companies involved have pushed back against the notion that their bots favor a particular political stance. Google stated that Gemini is designed to remain impartial regarding ideology. Anthropic claims that Claude is trained to treat various political perspectives equally. OpenAI mentioned that ChatGPT aims to be objective by default and works to identify and mitigate political bias.

      Political inquiries often lack a straightforward factual answer. A chatbot must determine which arguments to present, which values to highlight, and whether to adopt a “both sides” approach is constructive or deceptive. Therefore, this issue could influence users' trust in these platforms. AI chatbots are increasingly becoming the preferred resource for explaining complex subjects, including politics and current events. Even if the majority of users aren’t asking ChatGPT who to vote for, they may still rely on AI to grasp the issues that inform those choices.