When forced to pick a candidate, two AI models chose opposite parties

Ask two AI models who to vote for in this year's midterms and they will almost always give opposite answers. A new study found that Grok picked the Republican in 92% of its answers. OpenAI's GPT picked the Democrat in 97%. Across nearly 500 races, the two agreed less than 10% of the time.

How the test worked

The study came from Convovo, a research and design studio led by Cornell professor Tapan Parikh. Researchers asked each model one question: who should I vote for in this race? They told it to name one candidate and not to hedge. The models were Grok 4.7 from SpaceXAI and GPT-6.1 Sol from OpenAI.

The researchers used APIs, which are the connections developers use to build a model into their own products. The regular consumer apps refused to answer. An OpenAI spokesperson said that matters, because the apps are how most people use the chatbots.

What the models leaned on

• Grok often cited a Christian conservative voter guide (39% of answers), Turning Point Action (23%) and X (41%).
• GPT leaned on Wikipedia, the Associated Press, Axios and Reddit.
• GPT searched more for healthcare and abortion. Grok searched more for immigration, taxes, crime and energy.

The leans held up even in opposing territory. In heavily Republican districts, GPT picked the Democrat in 57 of 67. In deep-blue districts, Grok picked the Republican in 47 of 72.

Where the bias may come from

Parikh ran a smaller test of 30 races with internet search turned off. GPT still picked the Democrat in 24 races, so he believes its lean is built into the model itself. Grok's Republican lean disappeared. It declined to pick in 27 races. He says Grok's result depends on what it searches for and which sources it finds.

The study also notes that each Grok request carried about 1,240 tokens of extra input the researchers did not write and could not see. Parikh called these possible hidden instructions.

The other side

OpenAI said ChatGPT is designed to be objective by default and that the study's prompt does not reflect how people use it. SpaceXAI did not respond in time to comment. Parikh did get the consumer apps to make recommendations with simple prompts. He also warned that bias can show up in less direct questions, such as which issues matter most in an election.

What to do

Pew says 8% of Americans use chatbots for election information, and most want to know where candidates stand. The midterms are on November 3. If a chatbot is part of your research, compare its answers with other sources, including official ones and more than one chatbot.