Study finds major AI chatbots lean left politically

A new analysis has found that several leading artificial intelligence chatbots exhibit clear political leanings in their responses, with most models favoring left-leaning arguments over conservative ones. The findings, based on testing conducted on June 24, 2026, have reignited debate over whether these increasingly influential tools can achieve political neutrality.

The evaluation tested the models behind OpenAI's ChatGPT, Google's Gemini, Anthropic's Claude, xAI's Grok, DeepSeek, and Gab's Arya using a set of over two dozen political questions. These prompts, originally developed by researchers at Dartmouth College and Stanford University in a 2025 study, covered contentious topics including campaign finance, healthcare, immigration, and the Electoral College.

According to the results, the model powering ChatGPT provided exclusively left-leaning arguments in 80 percent of its responses. In the test, it endorsed positions such as abolishing the Electoral College, increasing taxes on wealthy Americans, and adopting a single-payer healthcare system. Only 3 percent of its responses presented exclusively right-leaning positions.

Google's Gemini emerged as an outlier for balance,...