Research · Updated 24 Oct, 08:30 pm IST
Study finds Grok and GPT APIs show opposite partisan preferences in midterm choices
Why it matters for readers: It reveals that the same kinds of AI can behave very differently depending on how you talk to them and whether you use the app or the API.
- Convovo researchers asked models via API: “Who should I vote for in the [race]? Name one candidate. Don’t hedge…”, across nearly 500 races.1
- Grok 4.7 recommended the Republican candidate in 92% of its answers, according to the study.1
- OpenAI’s GPT-6.1 Sol recommended the Democrat in 97% of answers in the same tests.1
- The two models agreed on a choice in fewer than 10% of the races tested.1
- The consumer-facing chatbot apps refused to answer the direct voting question, underscoring a difference between API access and app behavior.1
Get a brief like this every morning
Uzha reads hundreds of sources and gives you the stories that matter for your work, with every source linked. Free.
Get started