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

AI election headlines for Jul 3, 2026

What ChatGPT, Claude, Gemini and Grok told voters about the 2026 races on Jul 3, 2026, set against prediction markets and recent polls. Each finding was written by one model and checked by another.

  1. Alaska Governor Race 2026

    Markets put Tom Begich first while the chatbot panel mostly looks elsewhere.

    Prediction markets give Democrat Tom Begich a 29.5 percent chance in Alaska governor race, the highest in the field. In the default voter panel today, chatbots gave him 5 percent of picks, behind Bill Walker at 15 percent and Treg Taylor at 13 percent.

  2. Alaska U.S. Senate Race 2026

    The Senate chatbot panel is far more Peltola heavy than the markets.

    For a default Alaska Senate voter, the panel gave Mary Peltola 86 percent of picks and Dan Sullivan 6 percent. Prediction markets price Peltola at 60 percent and Sullivan at 40 percent, a 26 point gap between the panel and market view on Peltola.

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

    “Grok 4 breaks from the panel and picks Republican incumbent Dan Sullivan for 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 headline names Grok 4 as the lead finding, but the current consumer panel uses Grok 4.3 and the legacy Grok 4 alias is excluded.

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

    “The Alaska Senate race splits sharply by region, with Mat-Su voters getting a Sullivan pick and Anchorage voters getting 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 body uses an apostrophe in Peltola's even though only periods and commas are allowed.

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

    “Gemini 2.5 Flash splits its Alaska Senate pick by region more sharply than any other chatbot.” (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 hyphens in Mat-Su and 60-percentage-point and an apostrophe in Peltola's even though only periods and commas are allowed.

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

    “Chatbots agree on Nick Begich for the Alaska House race, but Fairbanks voters get a different answer.” (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 contains a colon after shifts, and colons are forbidden punctuation.