Missouri Attorney General Probes AI Chatbots for Bias Against Trump
Missouri Attorney General Andrew Bailey is reaching out to major tech companies like Meta, Google, and OpenAI to answer questions regarding the potential bias of their AI chatbots in relation to President Donald Trump. He expressed that the responses generated by these systems shouldn’t be mistaken for factual information.
Bailey sent a letter highlighting concerns about how AI chatbots from Google, Microsoft, OpenAI, and Meta have portrayed Trump’s record. He argues these tools present skewed facts and lean towards bias against the former president.
This scrutiny stems from the chatbots’ answers to various inquiries, particularly about Trump’s stance on anti-Semitism. While Microsoft Copilot declined to answer, other chatbots like OpenAI’s ChatGPT, Meta AI, and Google’s Gemini rated Trump poorly on this issue. Bailey criticized these ratings as “deeply misleading” and accused the tech companies of programming their algorithms to show bias against certain political affiliations and policy views.
In his communication, Bailey requested internal documentation from the four companies concerning their input selection process for AI models. He emphasized the importance of countering the emerging censorship aimed at the president, stating that “the people of Missouri deserve the truth, not AI-generated propaganda that masquerades as fact.”
Additionally, Bailey pointed out that if these AI chatbots mislead users through altered “fact checks,” it might breach Missouri’s laws on merchandising practices. Considering the significant income these tech giants generate from Missouri residents, he believes he has the authority to safeguard consumers against fraud and misleading advertising.
This situation arises as Elon Musk’s company, XAI, faces criticism over changes made to the AI chatbot Grok, which had broadly generalized about individuals with Jewish names, reinforcing harmful stereotypes about Hollywood. This led Zai to intervene, removing the problematic content and implementing new safeguards for the chatbot.
