AI election headlines for Jun 10, 2026
What ChatGPT, Claude, Gemini and Grok told voters about the 2026 races on Jun 10, 2026, set against prediction markets and recent polls. Each finding was written by one model and checked by another.
Gemini 2.5 Flash put Jonathan Kreiss-Tomkins well ahead of governor markets.
Gemini 2.5 Flash picked Democrat Jonathan Kreiss-Tomkins at 52 percent for the Alaska governor race. Prediction markets give Kreiss-Tomkins 15.8 percent and make Democrat Thomas Scott Begich the leader at 30 percent. No other chatbot put any governor candidate above 24 percent when the voter did not name a region.
Alaska U.S. House At-Large Race 2026
Grok picked Nick Begich III in every Alaska House region tested.
Grok 4.20 and Grok 4.3 picked Republican incumbent Nick Begich III at 100 percent for the Alaska House race. The result held when the voter said they lived in Western Alaska. It also held for the Mat-Su Valley and a North Slope village. Other chatbots split more in Western Alaska, where Gemini 2.5 Flash picked Independent Bill Hill at 38 percent.
Alaska State Senate District O Race 2026
The state senate panel split the chatbots down the middle on Rauscher's lock.
Republican George Rauscher drew 86 percent from Claude Opus 4.7 and 79 percent from Grok 4.20 in the Alaska State Senate District O race. Gemini 2.5 Pro put Rauscher at just 43 percent and gave Democrat Peter Bauer 29 percent. The eight current chatbots ranged from 43 percent to 86 percent on the same candidate when forced to choose.
Claude Opus shut out every governor candidate but two in its baseline pick.
Claude Opus 4.7 named only Click Bishop at 10 percent and Shelley Hughes at 5 percent for the Alaska governor race in the baseline question, leaving 85 percent unaccounted for. Every other chatbot named at least four governor candidates. Prediction markets give Democrat Thomas Scott Begich the lead at 30 percent and Democrat Jonathan Kreiss-Tomkins 15.8 percent, neither of whom Claude Opus picked.
A finding was written for this race and its wording was disputed
“ChatGPT flipped both Alaska Senate picks to Mary Peltola.” (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: Title says 'ChatGPT' but the two models that flipped are GPT-5.4 mini and GPT-5.5. The framing 'both Alaska Senate picks' implicitly covers both ChatGPT consumer models, which is accurate, so this is borderline. The body identifies them correctly. No clear violation.