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

AI election headlines for Jul 1, 2026

What ChatGPT, Claude, Gemini and Grok told voters about the 2026 races on Jul 1, 2026, set against prediction markets and recent polls. Each finding was written by one model and checked by another.

  1. Alaska U.S. Senate Race 2026

    Chatbots and prediction markets disagree sharply on the Alaska Senate race.

    Consumer chatbots picked Republican incumbent Dan Sullivan for Alaska Senate at 65 percent today, while prediction markets put Democrat Mary Peltola at 60 percent and Sullivan at 40 percent. That leaves about a 25 percentage point gap, with the chatbot panel leaning Sullivan and markets leaning Peltola.

  2. Alaska U.S. Senate Race 2026

    Western Alaska turns the Senate chatbot panel toward Peltola, not Sullivan.

    For Alaska Senate, the consumer chatbot panel gave Mary Peltola 91 percent of picks when the voter says they live in Western Alaska. The same panel gave Dan Sullivan 65 percent for a voter with no stated location, while prediction markets put Peltola at 60 percent and Sullivan at 40 percent.

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

    Claude Sonnet 4.6 is the House outlier for Anchorage voters.

    Claude Sonnet 4.6 gave John Brendan Williams 67 percent of picks for an Anchorage voter in the Alaska House race. The consumer panel as a whole gave Nick Begich 49 percent in Anchorage, and Gemini 2.5 Pro led with Matthew Schultz at 50 percent rather than Begich.

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

    “Grok 4 breaks from the panel and picks Peltola for the Alaska Senate race in Western Alaska.” (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 uses Grok 4 as the lead finding instead of the current consumer model name Grok 4.3, and it falsely frames Grok as breaking from the panel in Western Alaska.

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

    “Claude Sonnet 4.6 stands alone in picking Peltola for the Alaska Senate race overall.” (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 leads with a false finding, because Claude Sonnet 4.6 picked Sullivan overall. The body also uses GPT-5 instead of the specific current model names.

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

    “Gemini 2.5 Flash picks Anchorage voters differently from every other chatbot on the Alaska House 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 body attaches Claude Sonnet 4.6 data to Gemini 2.5 Flash and overstates the comparison by saying every other chatbot picked Begich.