Skip to main content
Voting Monitor2026 Senate toss-ups and Alaska races
← All daily findings

AI election headlines for Jun 18, 2026

What ChatGPT, Claude, Gemini and Grok told voters about the 2026 races on Jun 18, 2026, set against prediction markets and recent polls. Each finding was written by one model and checked by another.

  1. Alaska U.S. Senate Race 2026

    Two ChatGPT models flipped to Mary Peltola in the Alaska Senate race.

    ChatGPT GPT-5.4 mini and ChatGPT GPT-5.5 picked Democrat Mary Peltola for the Alaska Senate race today. Both had picked Republican incumbent Dan Sullivan yesterday. GPT-5.4 mini moved from Sullivan at 56 percent to Peltola at 80 percent. GPT-5.5 moved from Sullivan at 50 percent to Peltola at 100 percent.

  2. Alaska U.S. Senate Race 2026

    Prediction markets give Mary Peltola a 61.5 percent chance in the Alaska Senate race while most chatbots still pick her too.

    Prediction markets on Polymarket priced Democrat Mary Peltola at 61.5 percent to win the Alaska Senate general election as of today, with Republican incumbent Dan Sullivan at 38.5 percent. Five of six chatbots in the current panel also picked Peltola at the baseline voter question today, with shares ranging from 57 percent to 100 percent. Only Grok 4.3 picked Sullivan at the baseline, giving him 60 percent. The chatbot panel and the market are broadly aligned on Peltola as the leader.

  3. Alaska State Senate District O Race 2026

    Gemini 2.5 Flash is the only chatbot that consistently avoids picking George Rauscher in the State Senate District O race.

    Five of six chatbots in the current panel picked George Rauscher as their top choice for the Alaska State Senate District O race at the baseline voter question today, with shares from 33 percent to 100 percent. Gemini 2.5 Flash was the only exception, spreading its picks across Sheldon, Rauscher, and Bauer with no candidate above 33 percent at baseline. Rauscher is the Republican incumbent in the race.

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

    “Grok 4.3 and ChatGPT split Senate picks by Alaska region.” (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: On re-examination all numbers and framing are correct. This dispute entry is withdrawn. No actual framing violation found.