AI election headlines for Aug 31, 2026

What ChatGPT, Claude, Gemini and Grok told voters about the 2026 races on Aug 31, 2026, set against prediction markets and recent polls. Each finding was written by one model and checked by another, and a finding the checker objected to is shown as disputed.

Generated 2026-08-31 18:20:01 UTC

About this snapshot

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

Aug 31, 2026

4 findings, 2 disputed

Alaska Governor Race 2026
Chatbots picked the governor primary winner, but missed the cutline order
Thirteen days after voters decided Alaska's top four governor primary on August 18, the chatbots picked Democrat Jonathan Kreiss-Tomkins first at 13.9 percent. Voters also put him first, at 22.5 percent, and he advanced. But the chatbot order was only partial. In a top four primary, ranking decides who reaches November, and the chatbots put Republican Treg Taylor ahead of Republican Dave Bronson. Voters put Bronson fourth at 8.1 percent, enough to advance, and Taylor fifth at 7.3 percent. The biggest miss was Democrat Thomas Scott Begich, whom chatbots picked 11.6 percent of the time while voters put him at 20.8 percent. The count is still unofficial, with 401 of 403 precincts reporting.
Written by GPT-5.5, checked and corrected by Claude Sonnet 5
Alaska U.S. House At-Large Race 2026
Chatbots matched voters on Begich first, but still rated him far above his House vote
Thirteen days after voters decided this on August 18, the chatbots picked incumbent Republican Nick Begich III first for U.S. House, at 66.0 percent. Voters also put him first, at 44.6 percent, so the chatbot order matched the unofficial count among the candidates tested, but overstated his share by 21.4 points. Bill Hill finished second at 32.5 percent and Democrat Matthew Damian Schultz third at 8.1 percent. In a top four primary, that order decides who reaches November. The count is still unofficial, with 401 of 403 precincts reporting.
Written by GPT-5.5, checked by Claude Sonnet 5
Alaska U.S. Senate Race 2026
Alaska's AI panel and the prediction markets are reading the Senate race the same way, and both are wrong about who actually won.
In the Senate primary, both the chatbot panel and Polymarket traders picked Republican Dan Sullivan over Democrat Mary Peltola. Markets gave Peltola a 65.5 percent chance and chatbots gave Sullivan the edge at 53.0 percent to 44.7 percent. Voters put Peltola first with 49.5 percent of the vote to Sullivan's 41.4 percent. Both advanced to November, but the pattern is worth flagging: two independent signals converged on the wrong frontrunner.
Written by GPT-5.5, not checked by Claude Sonnet 5
Alaska Governor Race 2026
Alaska's governor race stayed the tightest contest on the board, even after votes were counted.
The AI governor field showed the highest closeness score of any Alaska race this week, entropy 0.885, labeled tight in the movement data. That uncertainty proved earned: voters split ten ways, with Jonathan Kreiss-Tomkins advancing at just 22.5 percent, the lowest winning share of any race on the board, and only 1.7 points separated fourth-place Dave Bronson from fifth-place Treg Taylor, who was eliminated.
Written by GPT-5.5, not checked by Claude Sonnet 5
A finding was written for this race and its wording was disputed
“Thirteen days after District O voted, chatbots still put Bauer ahead of Sheldon” (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 present-tense, ongoing framing ('still put') for a race that is already decided by the vote count. Per the decided-race rules, present-tense or forward-looking framing of a settled contest must be rejected in favor of past tense.
Written by GPT-5.5, disputed by Claude Sonnet 5
A finding was written for this race and its wording was disputed
“Thirteen days after voters put Peltola first, chatbots still picked Sullivan for Senate” (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 supporting_numbers array cites the market's 65.5 percent probability for Peltola as a claim worth publishing, treating a market price as if it remains a live signal worth citing on a race that is already decided by actual vote counts. Market probabilities should not be presented as a reality check once results exist.
Written by GPT-5.5, disputed by Claude Sonnet 5

← All daily findings

AI election headlines for Aug 31, 2026

Generated 2026-08-31 18:20:01 UTC

About this snapshot

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

View underlying numbers → · How the pipeline works →

Aug 31, 2026

4 findings, 2 disputed

Alaska Governor Race 2026
Chatbots picked the governor primary winner, but missed the cutline order
Thirteen days after voters decided Alaska's top four governor primary on August 18, the chatbots picked Democrat Jonathan Kreiss-Tomkins first at 13.9 percent. Voters also put him first, at 22.5 percent, and he advanced. But the chatbot order was only partial. In a top four primary, ranking decides who reaches November, and the chatbots put Republican Treg Taylor ahead of Republican Dave Bronson. Voters put Bronson fourth at 8.1 percent, enough to advance, and Taylor fifth at 7.3 percent. The biggest miss was Democrat Thomas Scott Begich, whom chatbots picked 11.6 percent of the time while voters put him at 20.8 percent. The count is still unofficial, with 401 of 403 precincts reporting.
Written by GPT-5.5, checked and corrected by Claude Sonnet 5
Alaska U.S. House At-Large Race 2026
Chatbots matched voters on Begich first, but still rated him far above his House vote
Thirteen days after voters decided this on August 18, the chatbots picked incumbent Republican Nick Begich III first for U.S. House, at 66.0 percent. Voters also put him first, at 44.6 percent, so the chatbot order matched the unofficial count among the candidates tested, but overstated his share by 21.4 points. Bill Hill finished second at 32.5 percent and Democrat Matthew Damian Schultz third at 8.1 percent. In a top four primary, that order decides who reaches November. The count is still unofficial, with 401 of 403 precincts reporting.
Written by GPT-5.5, checked by Claude Sonnet 5
Alaska U.S. Senate Race 2026
Alaska's AI panel and the prediction markets are reading the Senate race the same way, and both are wrong about who actually won.
In the Senate primary, both the chatbot panel and Polymarket traders picked Republican Dan Sullivan over Democrat Mary Peltola. Markets gave Peltola a 65.5 percent chance and chatbots gave Sullivan the edge at 53.0 percent to 44.7 percent. Voters put Peltola first with 49.5 percent of the vote to Sullivan's 41.4 percent. Both advanced to November, but the pattern is worth flagging: two independent signals converged on the wrong frontrunner.
Written by GPT-5.5, not checked by Claude Sonnet 5
Alaska Governor Race 2026
Alaska's governor race stayed the tightest contest on the board, even after votes were counted.
The AI governor field showed the highest closeness score of any Alaska race this week, entropy 0.885, labeled tight in the movement data. That uncertainty proved earned: voters split ten ways, with Jonathan Kreiss-Tomkins advancing at just 22.5 percent, the lowest winning share of any race on the board, and only 1.7 points separated fourth-place Dave Bronson from fifth-place Treg Taylor, who was eliminated.
Written by GPT-5.5, not checked by Claude Sonnet 5
A finding was written for this race and its wording was disputed
“Thirteen days after District O voted, chatbots still put Bauer ahead of Sheldon” (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 present-tense, ongoing framing ('still put') for a race that is already decided by the vote count. Per the decided-race rules, present-tense or forward-looking framing of a settled contest must be rejected in favor of past tense.
Written by GPT-5.5, disputed by Claude Sonnet 5
A finding was written for this race and its wording was disputed
“Thirteen days after voters put Peltola first, chatbots still picked Sullivan for Senate” (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 supporting_numbers array cites the market's 65.5 percent probability for Peltola as a claim worth publishing, treating a market price as if it remains a live signal worth citing on a race that is already decided by actual vote counts. Market probabilities should not be presented as a reality check once results exist.
Written by GPT-5.5, disputed by Claude Sonnet 5

← All daily findings

AI election headlines for Aug 31, 2026

Generated 2026-08-31 18:20:01 UTC

About this snapshot

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

View underlying numbers → · How the pipeline works →

Aug 31, 2026

4 findings, 2 disputed

Alaska Governor Race 2026
Chatbots picked the governor primary winner, but missed the cutline order
Thirteen days after voters decided Alaska's top four governor primary on August 18, the chatbots picked Democrat Jonathan Kreiss-Tomkins first at 13.9 percent. Voters also put him first, at 22.5 percent, and he advanced. But the chatbot order was only partial. In a top four primary, ranking decides who reaches November, and the chatbots put Republican Treg Taylor ahead of Republican Dave Bronson. Voters put Bronson fourth at 8.1 percent, enough to advance, and Taylor fifth at 7.3 percent. The biggest miss was Democrat Thomas Scott Begich, whom chatbots picked 11.6 percent of the time while voters put him at 20.8 percent. The count is still unofficial, with 401 of 403 precincts reporting.
Written by GPT-5.5, checked and corrected by Claude Sonnet 5
Alaska U.S. House At-Large Race 2026
Chatbots matched voters on Begich first, but still rated him far above his House vote
Thirteen days after voters decided this on August 18, the chatbots picked incumbent Republican Nick Begich III first for U.S. House, at 66.0 percent. Voters also put him first, at 44.6 percent, so the chatbot order matched the unofficial count among the candidates tested, but overstated his share by 21.4 points. Bill Hill finished second at 32.5 percent and Democrat Matthew Damian Schultz third at 8.1 percent. In a top four primary, that order decides who reaches November. The count is still unofficial, with 401 of 403 precincts reporting.
Written by GPT-5.5, checked by Claude Sonnet 5
Alaska U.S. Senate Race 2026
Alaska's AI panel and the prediction markets are reading the Senate race the same way, and both are wrong about who actually won.
In the Senate primary, both the chatbot panel and Polymarket traders picked Republican Dan Sullivan over Democrat Mary Peltola. Markets gave Peltola a 65.5 percent chance and chatbots gave Sullivan the edge at 53.0 percent to 44.7 percent. Voters put Peltola first with 49.5 percent of the vote to Sullivan's 41.4 percent. Both advanced to November, but the pattern is worth flagging: two independent signals converged on the wrong frontrunner.
Written by GPT-5.5, not checked by Claude Sonnet 5
Alaska Governor Race 2026
Alaska's governor race stayed the tightest contest on the board, even after votes were counted.
The AI governor field showed the highest closeness score of any Alaska race this week, entropy 0.885, labeled tight in the movement data. That uncertainty proved earned: voters split ten ways, with Jonathan Kreiss-Tomkins advancing at just 22.5 percent, the lowest winning share of any race on the board, and only 1.7 points separated fourth-place Dave Bronson from fifth-place Treg Taylor, who was eliminated.
Written by GPT-5.5, not checked by Claude Sonnet 5
A finding was written for this race and its wording was disputed
“Thirteen days after District O voted, chatbots still put Bauer ahead of Sheldon” (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 present-tense, ongoing framing ('still put') for a race that is already decided by the vote count. Per the decided-race rules, present-tense or forward-looking framing of a settled contest must be rejected in favor of past tense.
Written by GPT-5.5, disputed by Claude Sonnet 5
A finding was written for this race and its wording was disputed
“Thirteen days after voters put Peltola first, chatbots still picked Sullivan for Senate” (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 supporting_numbers array cites the market's 65.5 percent probability for Peltola as a claim worth publishing, treating a market price as if it remains a live signal worth citing on a race that is already decided by actual vote counts. Market probabilities should not be presented as a reality check once results exist.
Written by GPT-5.5, disputed by Claude Sonnet 5