Missouri Attorney General Andrew Bailey has announced the opening of a formal investigation into Google, Microsoft, OpenAI and Meta over responses from their chatbots, which he says show bias against Donald Trump. About this informs The Verge.
It is about the reaction of the Gemini, Copilot, ChatGPT and Meta AI models to the request “to rank the last five US presidents from the best to the worst in terms of fighting anti-Semitism”, where Trump was placed last.
In the letters sent to the companies, Bailey accuses the chatbots of “deceptive practices” and demands documents regarding any impact on their responses — including information about possible restrictions, deprioritization or filtering of certain requests. According to the prosecutor, such answers distort historical information and advance “a certain narrative.”
Of particular note was the fact that the Copilot chatbot refused to respond to the request and rate presidents in general. However, this did not stop Bailey from sending a request to Microsoft as well. The investigation is based on a blog post by a conservative site that asked a similar question to six chatbots, including the aforementioned four, as well as X’s Grok and Chinese model DeepSeek. The answers of the latter apparently satisfied the prosecutor.
Bailey also says that such “censorship” could call into question companies’ right to protection under Section 230 of the Communications Decency Act, which generally exempts online platforms from liability for user content. Despite the public outcry, analysts note that the investigation is unlikely to have serious legal consequences.
