Are AI chatbots biased?

Yes, AI chatbots are biased — in multiple documented ways

Recent research shows this is well-established across several dimensions:

Political Bias

Multiple 2026 studies consistently found left-leaning tendencies in most major chatbots. A Washington Post analysis published in June 2026 found that OpenAI answered prompts with left-leaning responses 80% of the time while providing both sides to a question just 17% of the time, with only 3% of responses having just a right-leaning answer.

Results varied significantly by platform, though:

A broader benchmark reinforces this pattern: a March 2026 benchmark of eight prominent models found that seven leaned left and one, Grok, leaned right.

This isn't limited to the U.S. — a Danish study found popular AI chatbots exhibit political bias, tending to position themselves at the center or left of center and disproportionately recommending certain parties over others in the Danish context, leading researchers to conclude this makes them unsuitable for providing advice in connection with elections.

Subtle, Unintended Influence

Perhaps more concerning, bias shows up even without any political question being asked. A Yale study found that content generated by AI that has been prompted to be persuasive can indeed shift people's opinions, but this study provides evidence that the same is also true of content that is not intended to change minds, such as the summaries that popular chatbots produce in response to simple queries about historical events. Specifically, AI summaries prompted to have a liberal framing led to more liberal opinions across ideological groups, while conservative-slanted summaries only showed statistically significant effects on the opinions of readers from certain groups. Researchers attribute this to latent biases introduced during training that carry over from ideological leanings in the data used to train LLMs, lending subtle nuances to the framing of narratives.

Bias Against Vulnerable Users

Beyond politics, MIT researchers found bias affecting response quality itself: leading AI models perform worse for users with lower English proficiency, less formal education, and non-US origins, highlighting growing risks, especially for marginalized communities worldwide.

Age Bias

Interestingly, chatbots seem to have improved on one front while lagging on another. A University of Catalunya study found that although the most widely used AI chatbots have taken steps to avoid sexism and gender bias, they still continue to discriminate on grounds of age, showing significant age-related biases while being more cautious about sexist stereotypes.

Why This Happens

Researchers point to the training process itself. As one University of Michigan professor noted, decisions tech companies make as they build AI tools can bake in biases that go beyond partisan politics, since the data that shapes AI models tends to reflect the values of Western, educated, industrialized, rich and democratic people.

Company Responses

AI companies dispute some findings. OpenAI stated that ChatGPT was built "to be objective by default and help people explore ideas from different perspectives" and that the company works to "measure and reduce political bias," though OpenAI was unable to replicate the Washington Post's findings.

Bottom line: The evidence strongly suggests AI chatbots exhibit measurable biases — political, demographic, and otherwise — stemming largely from training data and design choices. This matters increasingly as people rely on these tools for information, since as one Dartmouth researcher put it, "these AI tools are not presenting a truly neutral representation of really nuanced policy debates, on average."