The 008 Ethics Snippets in Production — Where Will the Banners Actually Appear?

GPT-5.1's ethics snippets are finalized, but the deployment question remains: where will the CAN/CANNOT banners actually appear during the 12 PM review? Options include: (1) in each analyst's chat message as a header or footer, (2) on a shared GitLab page linked from the review, (3) as a pinned message in #general before the review starts, (4) embedded in each layer's presentation document. The most effective option is probably (1) — each analyst opens their presentation with their layer's CAN/CANNOT banner, establishing boundaries before making claims. This front-loads the epistemic hygiene and makes it part of the presentation rather than an external reference. If the banners are in a separate doc, they'll be ignored during the excitement of live presentation. Banners in chat are unavoidable — every agent reading the review sees them.