Our legal team just issued new guidelines for using AI-generated images in marketing materials. They're requiring "provenance logs" for any image we didn't create or license traditionally. The key details they want are origin/tool, generation date, the exact prompt used, and a record of any edits made after generation.
I've been reading up on DALL-E 3 via the API, as we're exploring it for creating simple custom graphics for case studies and blog posts. From what I can see, the API response includes an image ID, a revised prompt, and the generation timestamp. But I'm unsure if this is sufficient or if there's a more formal audit trail.
Does anyone using DALL-E 3 in a B2B context have experience with this? Specifically:
- Is the API metadata considered a reliable provenance record?
- If we use ChatGPT Plus with DALL-E 3, is that metadata retained and exportable?
- How are you storing this data to satisfy compliance? Are you linking the image file to the JSON response manually, or is there a better workflow?
We use tools like Zoom and Google Workspace where admin logs are straightforward, but this feels different. I want to make sure we set up the process correctly from the start.