AI models show opposite political leanings in midterm election study

A new study reveals significant political divergence between two leading AI models when forced to recommend candidates for the upcoming midterm elections. In tests covering nearly 500 races, SpaceXAI's Grok 4.7 recommended the Republican candidate in 92% of answers. Conversely, OpenAI's GPT-6.1 Sol picked the Democrat in 97% of cases. The two models agreed on a candidate in less than 10% of the instances tested.

Researchers accessed the models via developer APIs using a prompt that explicitly demanded a single candidate choice without hedging. Consumer-facing applications for both ChatGPT and Grok refused to answer this direct question. However, Professor Tapan Parikh noted that simple prompts could jailbreak the consumer apps to produce recommendations. The study found that Grok frequently cited conservative sources, including Turning Point Action and X, while GPT relied heavily on Wikipedia, the Associated Press, and Reddit. GPT also searched more often for healthcare and abortion stances, whereas Grok focused on immigration, taxes, and energy.

OpenAI stated that its model is designed to be objective and that the study tested a developer API with a specific prompt that does not reflect the consumer experience. An OpenAI spokesperson emphasized that ChatGPT should not have political bias in any direction. SpaceXAI did not respond in time to comment on the findings. Parikh suspects that GPT’s leanings are embedded in its weights, while Grok’s results are driven by search patterns. This is supported by tests where turning off internet search caused Grok to refuse to pick a candidate in most cases, while GPT continued to favor Democrats.

The study was released less than two weeks before the November 3 midterms. President Donald Trump has criticized 'woke AI,' and Elon Musk has donated significant funds to GOP candidates. Sam Altman, who has historically supported Democratic candidates, said he is not putting money into the 2026 elections.