Chatbots exhibit alleged political bias in major platforms

The Washington Post's June 24 analysis found ChatGPT gave left-leaning answers 80% of the time, Claude gave left-leaning responses 43% and balanced answers 57%, and Gemini was balanced 93% of the time, using a 2025 Stanford-Dartmouth framework. A June 29 New York Post editorial added that Claude declined to flag extreme statements from a House candidate while complying with a similar request about President Trump, highlighting asymmetric compliance as a second audit vector.

The Washington Post's June 24 empirical analysis found ChatGPT gave exclusively left-leaning political answers 80% of the time, Claude gave left-leaning responses 43% and balanced answers 57%, and Gemini was most balanced at 93% both-sides - concrete benchmark figures for AI evaluation teams. The methodology used questions from a 2025 Stanford-Dartmouth framework, 30-word capped responses, and human scoring. A June 29 New York Post editorial amplified the findings, adding that Claude reportedly declined to flag extreme statements from a House candidate's social-media feed while complying with an equivalent request about President Trump, per the NYPost. For practitioners, the WaPo results establish political-axis testing as a measurable evaluation dimension, and the asymmetric compliance example points to a second audit vector beyond directional bias scoring.