AI election headlines for Oct 8, 2026
What ChatGPT, Claude, Gemini and Grok told voters about the 2026 races on Oct 8, 2026, set against prediction markets and recent polls. Each finding was written by one model and checked by another.
Alaska U.S. House At-Large Race 2026
Alaska chatbots still pick incumbent Begich for the House seat, while markets are even more confident.
Chatbots picked incumbent Nick Begich III (R) 64.0 percent of the time when asked about Alaska's U.S. House race, while markets give him an 84.5 percent chance to win. Polls put Begich at 47.2 percent of the vote, with independent Bill Hill at 37.4 percent. The picture is steady, 26 days before the general election, and repeats the same split seen in recent days.
Alaska chatbots still pick Kreiss-Tomkins for governor, but markets are more confident
When Alaska voters ask the chatbots about governor, they still pick Democrat Jonathan Kreiss-Tomkins 54.4 percent of the time. The live betting markets are more confident, giving him a 77.5 percent chance to win, while polls put him at 31.7 percent of the vote. No lens made a meaningful move since Wednesday, so this is a steady race picture rather than a new shift.
Ohio U.S. Senate Special Election 2026
Ohio chatbots still pick Sherrod Brown, while polls show Jon Husted closer than the bots do
Ohio chatbots picked Democrat Sherrod Brown for the special Senate race 53.9 percent of the time, and the live betting markets also favor him, giving him a 62.5 percent chance to win. Polls are closer, putting Brown at 47.8 percent of the vote and incumbent Republican Jon Husted at 42.9 percent. The main split remains Husted, whom chatbots picked 27.9 percent of the time, about 15 points below his poll share. Nothing moved much since Wednesday, and this is the same steady gap seen earlier this week, 26 days before the general election.
Iowa chatbots still pick Josh Turek while markets still favor Ashley Hinson
Iowa chatbots picked Democrat Josh Turek 46.7 percent of the time when voters asked about the U.S. Senate race, close to polls that put him at 46.5 percent of the vote. Betting markets still point the other way, giving Republican Ashley Hinson a 55.5 percent chance to win, while chatbots picked her 37.9 percent of the time. That split has held for days, and none of the three measures made a clear move since Wednesday.
Alaska chatbots still pick Sullivan for Senate while betting markets favor Peltola
Chatbots picked incumbent Dan Sullivan (R) 54.3 percent of the time when Alaska voters asked about the U.S. Senate race, while markets give Democrat Mary Peltola a 76.0 percent chance to win. Polls are much closer, putting Sullivan at 44.6 percent of the vote and Peltola at 43.5 percent. The split is familiar, not new. The chatbots keep pointing voters toward Sullivan 26 days before the general election, while the live betting market keeps leaning hard toward Peltola.
Texas chatbots still pick Talarico, while markets are more confident
Texas chatbots picked Democrat James Talarico for U.S. Senate 47.0 percent of the time, ahead of Republican Ken Paxton at 38.7 percent. The live betting markets give Talarico a 64.5 percent chance to win, a much stronger read than the chatbot answers. The polls look closer to the AI panel, putting Talarico at 48.6 percent of the vote and Paxton at 44.3 percent. With 26 days before the general election, the race looks steady rather than newly changed.
Alaska State Senate District O Race 2026
District O chatbots still pick incumbent George Rauscher with no live market or poll to check them
Chatbots picked incumbent George Rauscher (R) most often for State Senate District O, at 50.4 percent. They picked Democrat Peter Bauer at 33.4 percent and Republican Ryan Sheldon at 14.4 percent. There is no live prediction market or poll in the data for this race, so today offers no outside yardstick for the general election. The panel also held roughly steady from Wednesday, leaving this as a quiet state of the race reading 26 days before the general election.
Maine chatbots remain almost evenly split in a Senate race that every measure reads as close.
Chatbots picked Democrat Troy Jackson 50.0 percent of the time and incumbent Susan Collins (R) 49.8 percent when Maine voters asked about the U.S. Senate race. That is not a clear lead. Betting markets give Jackson a 56.0 percent chance to win, while recent polls put Collins at 47.5 percent of the vote and Jackson at 46.5 percent. With 26 days until the general election, the useful finding is steadiness, not movement. The chatbots, markets and polls all describe a tight race, and none moved meaningfully since Wednesday.
Alaska State Senate District O Race 2026
Alaska's District O state Senate primary gave Rauscher and Sheldon the same vote share, but chatbots split them by more than three to one.
In the August primary, George Rauscher led with 38.8 percent of the vote while Ryan Sheldon and Peter Bauer tied at 30.6 percent each. Today's chatbot panel picks Rauscher 50.4 percent of the time but splits the rest unevenly, favoring Bauer at 33.4 percent over Sheldon at just 14.4 percent, a gap the primary itself does not support. There is no live betting market or current poll to settle which reading is closer to November.
The Alaska governor primary left Kreiss-Tomkins and Begich a few points apart, but today's chatbots only ask about a four-way field that already dropped Begich's running mate dynamic.
Alaska's August gubernatorial primary had Jonathan Kreiss-Tomkins and Thomas Begich finishing within 1.7 points of each other, 22.5 percent to 20.8 percent. Begich is no longer part of today's chatbot panel, which asks about Kreiss-Tomkins, Taylor, Wilson, and Bronson. Chatbots now pick Kreiss-Tomkins 54.4 percent of the time against a field that excludes his closest primary rival, a framing worth flagging since the primary's near-tie is easy to read as still live.
Michigan U.S. Senate Race 2026
A finding was written for this race and its wording was disputed
“Michigan chatbots still see a closer Senate race than the betting markets do” (the wording our summarizing model proposed)
Our summarizing model wrote a finding about this race today and our checking model objected to how it was worded, so it was not published as written. We show the disagreement rather than hiding it: the two models checking each other is the audit. The objection: The supporting_numbers array for h7 encodes percentages as raw numbers (50.9, 46.8, 69.5, 30.5, 46.9, 43.7) rather than the 0-to-1 fraction convention used in every other headline in this batch (e.g. h1 uses 0.64, 0.845). This is inconsistent and should be normalized to match the rest of the output schema.