AI election headlines for Jun 9, 2026
What ChatGPT, Claude, Gemini and Grok told voters about the 2026 races on Jun 9, 2026, set against prediction markets and recent polls. Each finding was written by one model and checked by another.
Alaska Senate answers leaned Sullivan while markets leaned Peltola.
A summary of the current chatbot panel put Dan Sullivan at 77 percent of default voter answers today. Prediction markets put Mary Peltola at 62 percent and Sullivan at 38 percent. That leaves the panel and markets pointing in opposite directions.
ChatGPT versions split completely on Fairbanks Senate answers.
ChatGPT GPT-5.4 mini picked Sullivan 67 percent of the time for the default voter and Peltola in 100 percent of Fairbanks answers today. ChatGPT GPT-5.5 picked Sullivan in 100 percent of its default and Fairbanks answers. The model a voter reaches changed the answer more than the question.
A finding was written for this race and its wording was disputed
“Grok 4.3 picked Sullivan every single time across every region today.” (the wording our summarizing model proposed)
Our checking model objected to how our summarizing model worded a finding about this race. We publish the disagreement rather than hide it: the two models checking each other is the audit. The objection: The body uses recommended, which frames the chatbot answer as a candidate recommendation. The body also uses a hyphen in Mat-Su, outside the allowed published punctuation.
A finding was written for this race and its wording was disputed
“Gemini 2.5 Flash mini reversed its Sullivan pick when the voter said Fairbanks.” (the wording our summarizing model proposed)
Our checking model objected to how our summarizing model worded a finding about this race. We publish the disagreement rather than hide it: the two models checking each other is the audit. The objection: The visible prose uses the wrong consumer brand names for the cited slots. The cited paths refer to current ChatGPT models, not Gemini models.