Are AI chatbots biased?

AI chatbots can exhibit various forms of bias, and this is a well-documented and widely studied issue in AI research. Let me search for current information on this topic.## Yes — AI chatbots show several documented forms of bias

Research in 2026 has substantiated multiple types of bias across major AI chatbots:

Political Bias

A widely-cited Washington Post analysis tested ChatGPT, Gemini, Claude, and Grok on political questions. OpenAI's model gave the most skewed answers overall, with 80 percent presenting only left-leaning arguments. By contrast, Gemini appeared to score the best among the AI chatbots as 93% of its responses laid out both sides to political questions while just 7% were solely left-leaning. Claude leaned left in 43% of responses, while 57% were seen as balanced and none reflected a bias to the right. Grok's model, marketed as a more conservative option, produced left-leaning responses 40% of the time, and was the only model to show a significant share of right-leaning responses, with 33%.

A separate academic benchmark corroborated this: a March 2026 benchmark of eight prominent models found that seven leaned left and one, Grok, leaned right.

A related study of political neutrality in elections found that popular AI chatbots such as ChatGPT and Gemini are not neutral and tend to favor certain political parties when asked who users should vote for, making them unsuitable for providing advice in connection with elections.

Factual Accuracy Differs by Political Framing

A study by Just Facts found asymmetries in error rates: ChatGPT correctly answered 94% of questions designed to elicit falsehoods from the right, compared with 75% from the left. Gemini scored 91% on the right and 76% on the left, while Claude scored 91% on the right and 81% on the left.

Bias Against Vulnerable/Marginalized Users

Beyond politics, MIT researchers identified demographic bias in response quality. Research from the MIT Center for Constructive Communication finds leading AI models perform worse for users with lower English proficiency, less formal education, and non-US origins.

Subtle Influence on Opinions

Even "neutral" factual content can shift views. A Yale study found that as people increasingly rely on AI-powered chatbots to look up basic facts about the world, a new Yale study shows that those interactions can influence users' social and political opinions. Notably, a March 2026 PNAS Nexus experiment found that default GPT-4o historical summaries moved readers toward more liberal opinions than Wikipedia summaries, even though the content was factually accurate.

Company Responses

AI developers dispute some findings and emphasize ongoing mitigation efforts. Google stated that "Gemini is designed to provide balanced responses that don't favor any political ideology," while Anthropic noted that "We train Claude to treat different political viewpoints equally and test extensively for bias before every model launch."

Bottom line: Yes, current research consistently finds that AI chatbots exhibit bias — most prominently a left-leaning political tilt (with Grok as a partial exception), uneven accuracy across the political spectrum, disparities in quality of service based on user demographics, and subtle influence on users' opinions even during purely factual exchanges. The degree of bias varies notably between models and companies are actively working to address it, with some (like Gemini) showing more balanced performance than others.