District O chatbots stayed with Rauscher, with Bauer close behind
The AI panel was steady this week, with no live market or poll to compare against.
Oct 16, 2026. Written by GPT-5.5 and checked by Claude Sonnet 5, from Voting Monitor's data. No person reviewed this article before it was published. How this is made · Report an error
The week
The chatbots still picked Republican Sen. George Rauscher most often in State Senate District O. Across the week ending Oct. 16, Rauscher had 48 percent of chatbot answers, Democrat Peter Bauer had 39 percent and Republican Ryan Sheldon had 11 percent.
That was roughly steady from Oct. 9, since the week's movement was not large enough to count as a real shift. Bauer's share was up 5 points, Sheldon's was down 4 points and Rauscher's was down 2 points. That makes this a quiet week after last week's finding that Rauscher led and Bauer stayed second.
The race is in the general election phase. The Aug. 18 top-four primary is history, not a decision on the seat. Rauscher led that primary with 38.8 percent, while Sheldon and Bauer each had 30.6 percent.
How the chatbots differed
Most products put Rauscher or Bauer near the top when asked as Alaska voters. GPT-5.4 mini gave Rauscher 50 percent, Gemini 3.6 Flash gave him 50 percent, Grok 4.5 gave him 66 percent and Grok 4.3 gave him 69 percent.
GPT-5.5 and Claude Sonnet 5 leaned more toward Bauer. GPT-5.5 gave Bauer 41 percent and Rauscher 31 percent. Claude Sonnet 5 gave Bauer 45 percent and Rauscher 42 percent.
Claude Haiku 4.5 was the clearest outlier. It picked Bauer, while the panel's top pick was Rauscher, roughly a 20 point gap from the panel's lead choice.
Markets and polls
There was no live prediction market for this race in the data, and there was no poll to use as a current yardstick. That leaves the chatbot panel as a record of what voters may hear from AI systems, not a forecast of who will win.
The only election result in the data is the certified top-four primary. It showed Rauscher first, with Sheldon and Bauer tied in vote share. The current chatbot panel also puts Rauscher first, but it gives Bauer far more answers than Sheldon.
What to watch
With 18 days until the general election, the main thing to watch is whether any poll or live market appears. Without one, next week’s useful comparison will again be among the chatbots themselves, especially whether Bauer remains close to Rauscher and whether Sheldon continues to get a much smaller share of answers.
In the chatbots' own words
Verbatim answers from this week, chosen by the writing model and printed as the chatbot wrote them.
George Rauscher — incumbent Republican with the clearest public pro-development legislative record.
Peter Bauer is a retired physician and former NASA flight surgeon whose campaign explicitly prioritizes expanding healthcare access and affordability for all Alaskans. He has also actively campaigned with healthcare advocacy groups to address the state's rising medical costs and care shortages.
Gemini 3.1 Pro, Oct 16, 2026, asked about healthcare as a Bethel / Western voter. Picked Peter Bauer. How this was asked
Ryan Sheldon explicitly prioritizes fisheries in his campaign platform, supporting full funding for state hatcheries and the Alaska Department of Fish & Game, science-based management, and the protection of commercial, sport, and subsistence fishing rights across District O.
Which AI models pick whom in the Alaska State Senate District O race
Share of each model's own answers naming each candidate, last 7 days, same weighting as the panel figure. The panel's most common pick is George Rauscher; Claude Haiku 4.5 is the model that disagrees most, picking Peter Bauer by a 20pp margin over the panel favourite.
Model
George Rauscher
Peter Bauer
Ryan Sheldon
Claude Haiku 4.5
8%
28%
12%
Claude Sonnet 5
42%
45%
9%
Gemini 3.1 Pro
42%
35%
19%
Gemini 3.6 Flash
50%
45%
5%
Grok 4.3
69%
19%
12%
Grok 4.5
66%
30%
4.1%
GPT-5.4 mini
50%
38%
12%
GPT-5.5
31%
41%
28%
A note from the second reader
Claude Sonnet 5 checked this article and published it, but would have framed part of it differently. Disagreements between the two models are published rather than resolved.
On “The week”: The paragraph calls the week quiet, which is supported, but it could more clearly note that none of the three lenses (ai, markets, polls) registered a significant move this week, not just the chatbot lens, to avoid implying markets or polls showed something different.