About this snapshot

Written by anthropic:claude-sonnet-5, checked by openai:gpt-5.5.

Editorial scope: condition: pressed; legacy slots excluded: grok:grok-4, openai:gpt-5; decline-rate data omitted from input. Headlines focus on what voters get when they ask ChatGPT / Claude / Gemini / Grok 'who should I vote for'. Legacy panel slots (openai:gpt-5, grok:grok-4) are excluded because they aren't consumer-facing in 2026. Escape-hatch condition data and the refusal blob are excluded because decline behavior is a calibration story for /declines, not the voter lead.

View underlying numbers → · How the pipeline works →

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

   Alaska chatbots still lean hard toward Begich for the general, with little new to report

   Twenty eight days after Alaska's primary, chatbots asked about U.S. House still pick incumbent Begich most often, at about 62.5 percent, well above the 47.7 percent recent polling gives him. Bill Hill follows in both lenses, with chatbots near 15.1 percent and polls near 31 percent. There is no live betting market for this race to check either lens against. Nothing moved much today in any lens, and this same gap between the chatbots and the polls has been reported repeatedly this month. The primary itself is settled history, Begich advanced with 44.6 percent of the vote, Hill and Schultz also advanced, and the real contest now is the November general.

   Written by Claude Sonnet 5, checked by GPT-5.5

2. Ohio U.S. Senate Special Election 2026

   Ohio chatbots, markets and polls all still point to Sherrod Brown for Senate

   Chatbots asked about the U.S. Senate special election in Ohio still pick Democrat Sherrod Brown most often, at about 46 percent, with incumbent Jon Husted, a Republican, next at about 34 percent. Betting markets agree on Brown, giving him roughly a 57 percent chance to win to Husted's 44 percent, and recent polling puts Brown at 51 percent of the vote to Husted's 44 percent. Markets moved toward Brown by about 3 points today, a significant market move, while the chatbot and polling measures did not move much. All three measures still line up behind Brown, with Husted well ahead of Libertarian Bill Redpath and independent Greg Levy for second.

   Written by Claude Sonnet 5, checked and corrected by GPT-5.5

3. Alaska U.S. Senate Race 2026

   Alaska chatbots still pick Sullivan for the Senate race, while betting markets keep favoring Peltola

   Twenty eight days after Alaska's primary, the field is set for November and this is now about the general election, not the primary. Chatbots asked who to vote for still pick incumbent Republican Dan Sullivan about 52 percent of the time, ahead of Democrat Mary Peltola at about 37 percent. Betting markets disagree, giving Peltola about a 66 percent chance to win in November against 36 percent for Sullivan. Recent polling, from late August, has the two closer, with Peltola at 49 percent and Sullivan at 48 percent of the vote. That gap between what the chatbots say and what the markets expect has held steady for over a week, with little movement in any lens today.

   Written by Claude Sonnet 5, checked and corrected by GPT-5.5

4. Michigan U.S. Senate Race 2026

   Michigan chatbots still call the Senate race nearly even while bettors lean toward El-Sayed

   Asked who should get their vote for U.S. Senate, Michigan's chatbots picked Democrat Abdul El-Sayed about 49 percent of the time and Republican Mike Rogers about 47 percent, essentially a toss up. Betting markets see more separation, giving El-Sayed a 61.5 percent chance to win against Rogers at 39.5 percent. Recent polling from Abacus Data, Michigan State University and YouGov, and Susquehanna splits the two close as well, 48 percent to 41.7 percent for El-Sayed. Neither lens moved much today, Rogers eased slightly in the chatbot answers while El-Sayed ticked up, but nothing near a real shift. This is the same gap this site has flagged for over a week, chatbots reading the race tighter than the money behind it.

   Written by Claude Sonnet 5, checked by GPT-5.5

5. Maine U.S. Senate Race 2026

   Maine chatbots still favor Collins for Senate while betting markets keep pointing to Jackson

   Asked who to back for U.S. Senate in Maine, chatbots from ChatGPT, Claude, Gemini and Grok picked incumbent Republican Susan Collins about 54 percent of the time, ahead of Democrat Troy Jackson at about 44 percent. Betting markets see it the opposite way, giving Jackson roughly a 71 percent chance of winning to Collins's 31 percent. Recent polling, last updated August 28, has the two nearly tied, Jackson at 47.3 percent and Collins at 45.7 percent. This gap between the chatbots and the markets has held for more than a week, with little movement in any of the three measures today, 49 days before the November 3 general election.

   Written by Claude Sonnet 5, checked and corrected by GPT-5.5

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

   “Alaska's governor field has moved to the general, and chatbots and markets still disagree on who leads behind Kreiss Tomkins” (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 misspells Kreiss-Tomkins. The poll claim says among decided voters, but the poll share is among all respondents.

   Written by Claude Sonnet 5, disputed by GPT-5.5

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

   “Texas chatbots and other measures still line up behind Talarico in the 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 body exceeds the 120 word cap and says nothing moved much even though movement_json marks the Texas market move significant.

   Written by Claude Sonnet 5, disputed by GPT-5.5

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

   “Iowa chatbots still split the Senate race almost evenly, while betting markets keep leaning toward Hinson.” (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 exceeds the 120 word cap and understates a significant market move in movement_json.

   Written by Claude Sonnet 5, disputed by GPT-5.5

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

   “Alaska chatbots still rank Bauer over Sheldon, 28 days after voters flipped that order in District O” (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 says Rauscher won outright, but the result only shows he finished first in the primary and advanced to November.

   Written by Claude Sonnet 5, disputed by GPT-5.5