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

AI election headlines for Aug 2, 2026

What ChatGPT, Claude, Gemini and Grok told voters about the 2026 races on Aug 2, 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

    Chatbots still put incumbent Nick Begich III ahead in Alaska's House race, and there is no live market to check them against.

    Ask an AI chatbot who should win Alaska's U.S. House race and it points to Republican incumbent Nick Begich III about 66 percent of the time, above the 51 percent of the vote he pulls in recent polling. There is no live prediction market for this race, so that comparison cannot be made today. Polls from NYT Siena, Tavern Research and Alaska Survey Research, most recently from late June, put Democrat Matthew Damian Schultz at 37.3 percent, while chatbots give him just under 13 percent. Sixteen days remain before Alaska's top 4 primary, and nothing shifted enough today to call it real movement in any lens.

  2. Alaska State Senate District O Race 2026

    Chatbots still favor incumbent Rauscher in State Senate District O, with no market or poll to check the read.

    Ask a chatbot who should win Alaska's State Senate District O race, and it still names Republican incumbent George Rauscher most often, with the AI panel picking him about 43 percent of the time compared with 28 percent for Democrat Peter Bauer and 18 percent for Republican Ryan Sheldon. There is no live prediction market for this race and no poll data to check that lean against, so this remains an AI only read for now. Today's shift is small, Rauscher's AI share slipped a few points while Bauer and Sheldon each ticked up slightly, not enough to call a real change.

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

    “Alaska's betting markets just jumped toward Treg Taylor, 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 forbidden colons. It also exceeds the 120 word hard cap.

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

    “Chatbots still steer Alaska House voters to incumbent Begich, and there's no live market to check that 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 forbidden parentheses and a slash in NYT/Siena. It also adds unsupported causal speculation about why Schultz trails in chatbot picks.

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

    “Alaska's Senate race still splits the chatbots from the betting markets, and today the gap widened a bit as Peltola's market price ticked up.” (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 forbidden parentheses, colons and a slash in NYT/Siena. It also exceeds the 120 word hard cap.