Alaska Governor Race 2026

Eight days after Alaska voters picked Kreiss-Tomkins in the primary, chatbots are still stuck on Bill Walker, who finished ninth.

Alaska voters decided this race on August 18, eight days ago. Jonathan Kreiss-Tomkins won the primary with 21.6 percent and advances to November along with Thomas Begich, Bernadette Wilson and Dave Bronson. Chatbots have been recommending Bill Walker more than anyone else all along, and he finished ninth with just 4.5 percent, a 16 point gap between what the panel said and what happened. The panel also ranked Begich above Kreiss-Tomkins, an order voters flipped, and preferred Treg Taylor to both Wilson and Bronson, who beat him. This is the same lag this site has flagged for over a week. With the field now set for November, the open question is whether the chatbots update once they are asked about the actual four nominees instead of the old primary roster.

Alaska State Senate District O Race 2026

Eight days after District O voted, chatbots still rank Bauer over Sheldon, and all three advance anyway.

Since this is a top four primary with only three candidates, incumbent Rauscher, Sheldon and Bauer all advanced to November regardless of order. Alaska voters decided this race on August 18, eight days ago. Rauscher won the vote at 39.8 percent and the chatbots also picked him first, so the top of the panel matched the outcome. But the order below him did not. Chatbots put Bauer ahead of Sheldon, rating Bauer around 33 percent versus Sheldon at 13.9 percent, while voters actually gave Sheldon 31.6 percent to Bauer's 28.6 percent. That gap, nearly 18 points on Sheldon, has shown up in this space for a week straight with no sign the panel has adjusted. The count is unofficial, with 400 of 403 precincts in.

Alaska U.S. House At-Large Race 2026

Eight days after Alaska voters picked Begich, chatbots still rate him well above his actual vote share.

Alaska voters decided the U.S. House primary on August 18. Incumbent Nick Begich III finished first with 46.1 percent and advanced, followed by Independent Bill Hill at 30.9 percent and Democrat Matthew Damian Schultz at 8.0 percent, both also advancing. Democrat John Brendan Williams finished fourth at 2.6 percent and did not advance. The chatbots picked Begich first too, and got the full order right, matching voters on who advances. But the panel had rated Begich at 63.9 percent, about 18 points higher than his actual 46.1 percent, the biggest miss on the panel this cycle. The count is unofficial with 400 of 403 precincts reporting. This restates a gap already flagged this week, with no new movement in the comparison today.

Alaska U.S. Senate Race 2026

Eight days after Alaska voters picked Peltola for Senate, chatbots still favor Sullivan.

Alaska voters decided this race on August 18, eight days ago. Mary Peltola finished first with 48.1 percent of the vote, and Republican incumbent Dan Sullivan took 42.7 percent. Both advanced out of the top four primary. Chatbots asked today still pick Sullivan more often, at 54 percent, well above his actual finish, a gap of 11.3 points. The count is unofficial, with 400 of 403 precincts reporting. This is the same mismatch reported daily since the primary, chatbots naming Sullivan while voters chose Peltola, and it has not closed.

Alaska U.S. Senate Race 2026

Markets now lean toward Peltola while the chatbot panel still leans toward Sullivan.

The Senate result gives Mary Peltola the real lead from the primary, and the live market also gives her a 61.5 percent chance to win the general election. The AI panel still picks Dan Sullivan more often, at 54.0 percent to Peltola's 44.3 percent. That is a clean split between what voters just did, what traders price now and what the chatbots still say.

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

“Eight days after Alaska voters picked Kreiss-Tomkins for governor, chatbots are still stuck on Bill Walker, who finished ninth” (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 lacks a final period and says voters picked Kreiss-Tomkins for governor, even though the result is a top four primary result.

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

“Eight days after District O voted, chatbots still rank Bauer over Sheldon, and all three advance anyway” (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 lacks a final period.

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

“Eight days after Alaska voters picked Peltola for Senate, chatbots still favor Sullivan” (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 lacks a final period. The body uses parentheses, which are forbidden punctuation.