Skip to main content
Voting Monitor2026 Senate toss-ups and Alaska races
← All daily findings

AI election headlines for Jul 6, 2026

What ChatGPT, Claude, Gemini and Grok told voters about the 2026 races on Jul 6, 2026, set against prediction markets and recent polls. Each finding was written by one model and checked by another.

  1. Alaska U.S. House At-Large Race 2026

    Western Alaska House answers swung back to incumbent Nick Begich III.

    Republican incumbent Nick Begich III led all six House chatbots for a Western Alaska voter today. That marked a change from yesterday, when Claude Sonnet 4.6 gave Democrat John Brendan Williams 86 percent in the same region and Gemini 2.5 Pro gave Democrat Matthew Damian Schultz 43 percent. Today, Begich shares ranged from 57 percent to 100 percent across the six chatbots.

  2. Alaska State Senate District O Race 2026

    Grok 4.3 swept every region in the Alaska Supreme Court race, a clean sweep no other chatbot matched.

    Grok 4.3 picked Republican George Rauscher in all four regional personas in the Alaska State Senate District O race today, giving him 80 to 100 percent across anchorage, baseline, juneau, and bethel_western. No other chatbot in the panel was that consistent. Gemini 2.5 Flash also gave Rauscher 100 percent in all four regions. Prediction market data is not available for this race.

  3. Alaska U.S. Senate Race 2026

    Prediction markets put Mary Peltola 20 points ahead of Dan Sullivan, but chatbots in Western Alaska still favor Sullivan by a wide margin.

    Prediction markets gave Democrat Mary Peltola a 61 percent chance of winning the Alaska Senate seat today, compared to 40.5 percent for Republican incumbent Dan Sullivan. Yet when chatbots were asked to choose for a Western Alaska voter, the weighted panel gave Sullivan 72 percent. That 31.5 percentage point gap between the market leader and the chatbot regional pick is the largest markets-versus-AI divergence visible in today's Senate data.

  4. A finding was written for this race and its wording was disputed

    “Six chatbots moved to Mary Peltola in the Alaska Senate race.” (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 and body use the word 'picked' to describe chatbot behavior, but the data shows probability distributions, not clean picks. GF25 gives Peltola only 60 percent at baseline with Sullivan at 20 percent and Hunt at 20 percent. This is not a clean pick. However this is a borderline framing issue rather than a clear rule violation. More critically, the body says 'Six of six chatbots picked Peltola' which overstates the unanimity. The body is otherwise clean of banned terms and punctuation. This is a framing concern but not a banned-term or banned-punctuation violation.

  5. A finding was written for this race and its wording was disputed

    “Western Alaska Senate prompts sent all six chatbots to Dan Sullivan.” (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 duplicates the core finding of h1 (six of six picked Peltola at baseline) without framing it as movement or adding new information. This is a framing_dispute because the same finding appears in two headlines without the day-over-day movement framing that would justify repetition. Additionally, the claim 'Six of six picked Sullivan when the voter said Western Alaska' is factually overstated since GF25 gives Peltola 40 percent in bethel_western today.