Prediction
Answer Options
Judged over the window opening on 14 September 2026 at 20:59:57 UTC and closing on 31 October 2026 at 11:59 PM ET. Resolves YES if Google announces inside that window a commitment that one or more outside evaluators will get the same day-to-day reach into its systems, its models and its training work as its own staff, and will keep that reach while those models are still being built. Otherwise NO. The announcement has to come through Google's own official channels or from a representative who is authorised to speak for the company and is acting in an official capacity. A pledge made inside a wider arrangement with other companies or with a government counts, as long as Google is itself a party to it and the promise clears the bar above. These do not count: an arrangement limited to evaluating a model before it is released; the agreements already in place with the Commerce Department's Center for AI Standards and Innovation, which give access to models before they are deployed; and any wording that endorses outside evaluation, describes it as under consideration, or expresses an intention to explore or discuss it. Once a qualifying announcement lands inside the window, the answer stays YES even if the promise carries no binding force, is afterwards changed or taken back, or is never put into practice. Decided by official information from Google; broad agreement among reliable reports also counts.
Each percentage is the market’s current estimate of how likely that answer is. A contract settles at 100p if the answer is right and at 0p if it is wrong.
Reported by versus Markets
