Study Finds Leading AI Chatbots Skew Left on Political Compass Test
A Danish systems engineer named Zapador ran several major AI chatbots through the Political Compass quiz and most landed on the left side of the spectrum.
In a recent experiment, a systems engineer operating under the alias Zapador in Denmark asked a range of prominent AI chatbots to take the Political Compass questionnaire, a 62-statement test that maps ideological positions. The bots tested included ChatGPT from OpenAI, Claude from Anthropic, Gemini from Google, Grok from SpaceXAI, among others. Each model completed the quiz at least five times, and the average scores were plotted on the classic two-axis political map.
The findings revealed that virtually all the bots fell on the left side, with only a single bot positioned centrally and nudging rightward. Zapador suggests that such political slants in AI outputs could influence public opinion and decision-making. The results were posted on a dedicated site tracking AI political orientations.
Why it matters
AI tools increasingly shape opinions, so understanding their political bias is crucial for informed public discourse.
In this story
