Microsoft's Transparency Note for Azure OpenAI describes a face-blurring preprocessing step applied to image and video inputs. The relevant passage states:
For inputs to GPT-4 Turbo with Vision, GPT-4o, and GPT-4o mini that contain images or videos of people, the system will first blur faces prior to processing to return the requested results.
The document also explicitly notes two exceptions where blurring is not applied:
Computer Use (preview) "does not use face blurring, as it has the potential to obscure UI elements and degrade performance"
GPT-image-1 "does not use face blurring"
GPT-5 (and the broader GPT-5 series, including GPT-5.1-Codex-Max) supports vision inputs per the same document's capability matrix, but is not mentioned in either the blurring list or the non-blurring exceptions.
Does the face-blurring preprocessing step apply to GPT-5 vision inputs on Azure OpenAI?
Specifically:
Is blurring applied silently (and the doc just hasn't been updated, as it typical in Microsoft…