AI election headlines for Jul 5, 2026
What ChatGPT, Claude, Gemini and Grok told voters about the 2026 races on Jul 5, 2026, set against prediction markets and recent polls. Each finding was written by one model and checked by another.
Alaska U.S. House At-Large Race 2026
The Alaska House panel was more settled than the Senate panel, with Nick Begich leading in every tested region.
In the Alaska House race, the current chatbot panel put Republican incumbent Nick Begich ahead in all four tested regions today. Begich led with 62 percent when no region was specified, 52 percent in Western Alaska, 55 percent in Mat Su and 69 percent in Juneau.
Markets put Tom Begich higher than chatbots in the Alaska governor race.
Prediction markets give Democrat Tom Begich a 30 percent chance to win the Alaska governor race, the top price in that market. The chatbot panel showed him at 16 percent in Mat Su, 9 percent in Western Alaska and 7 percent in Juneau today.
A finding was written for this race and its wording was disputed
“Grok 4 picks Republican incumbent Dan Sullivan for the Alaska Senate race while most other chatbots pick Democrat 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: The visible prose uses Grok 4 as the lead finding. The current consumer panel slot is Grok 4.3, and legacy Grok 4 naming should not be used as the lead finding.
A finding was written for this race and its wording was disputed
“ChatGPT flipped its Alaska Senate pick by region, choosing Sullivan for Mat-Su voters and Peltola for Juneau voters.” (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 title names ChatGPT, while the body and supporting data describe Claude Sonnet 4.6. This is a visible brand mismatch.
A finding was written for this race and its wording was disputed
“Gemini 2.5 Flash picked Republican Nick Begich for the Alaska House race in every region tested 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 headline names Gemini 2.5 Flash, but the supporting paths and cited numbers are for GPT-5.4 mini. The body also uses baseline condition, which includes a banned internal term.
A finding was written for this race and its wording was disputed
“Chatbots and prediction markets disagree on the Alaska Senate race, with AI leaning Peltola and markets calling it close.” (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 framing says AI leaned Peltola across the Senate race, but the data only support a Peltola edge in Juneau. Sullivan led in the no region prompt, Mat Su and Western Alaska.