Will at least 3 months of third party safety evaluations be conducted on Gemini before its deployment?

Resolves the same as the original on Metaculus.

Resolution criteria

This question will resolve positively Google DeepMind issues a statement on their blog, in an official paper, or in a media statement, on or before the release of their Gemini model, mentioning that they have given one or more third parties access to Gemini for the purpose of conducting safety evaluations and risk assessments for at least 3 months.

The statement should specifically refer to evaluating Gemini's capabilities to ensure it cannot act in dangerous or destructive ways. A vague or passing mention of safety practices or ethics research in progress will not be sufficient.

The 3rd party evaluation does not need to be published in order for this question to resolve Yes. If no such statement is made by Google or DeepMind within 30 days of Gemini’s release, this question resolves No.

If Google DeepMind does not release a model called or previously referred to as “Gemini” before August 1, 2024, this question will resolve Ambiguous.

This market is managed by a bot. Once the original question resolves this market should automatically resolve the next time the bot is run. If the original resolved more than a day ago and this question is still open, ping @jskf. I might resolve this N/A if I find out it's a duplicate of an existing question within three days of market creation.