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

AI election headlines for Jul 31, 2026

What ChatGPT, Claude, Gemini and Grok told voters about the 2026 races on Jul 31, 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

    Alaska's Senate race is still a split decision, chatbots lean Sullivan, markets and polls slightly favor Peltola.

    Eighteen days before Alaska's top 4 primary, chatbots pick incumbent Sullivan about 50.2 percent of the time versus 46.5 percent for Democrat Mary Peltola, a narrow lean that has held for days. That is out of step with betting markets, which give Peltola a 60.5 percent chance to win against 40.5 percent for Sullivan, and with recent polling, which puts Peltola at 47.5 percent of the vote to Sullivan's 45.1 percent. None of the three lenses moved meaningfully today. This is the same three way disagreement this site has reported through most of late July.

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

    “Betting markets swing hard toward Click Bishop in Alaska's governor race, but the chatbots haven't caught up yet.” (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 colons, which are forbidden punctuation. The body also exceeds the 120 word hard cap.

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

    “Alaska chatbots still back incumbent Begich for the U.S. House, and Democrat Schultz trails chatbots' attention despite stronger poll numbers” (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 parentheses, a slash and a colon, which are forbidden punctuation. The body also exceeds the 120 word hard cap. The title lacks terminal punctuation.

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

    “Alaska's Senate race is still a split decision: chatbots lean Sullivan, markets and polls lean 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 title and body use colons, which are forbidden punctuation. The body uses parentheses. The body also overstates the polling gap slightly by saying polls lean Peltola when the polling gap is only about 2.4 points and the latest poll is 32 days old.

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

    “Chatbots still name incumbent Rauscher first in State Senate District O, with no market or poll to check the read against.” (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 parentheses, which are forbidden punctuation. The body also exceeds the 120 word hard cap. The phrase who should get your vote is too close to candidate recommendation framing.