If I can access a file, can I use it in an AI project?
Being able to open a file doesn't tell you what you're allowed to do with it. Reading it for a meeting, uploading it to a service, creating a derivative asset and publishing an excerpt are separate actions. Check each one with the responsible owner.
This comes up when a workflow moves material into a new tool or turns it into public content. Write down the source and why your team has access, then the proposed new use. Then check the relevant agreement, policy or permission with the responsible owner. An AI tool's ability to ingest the file doesn't answer the rights question.
If the scope is unclear, keep going where you can
Continue the work that doesn't depend on the reuse. Prepare a list of questions or use a cleared example, and hold only the affected part of the project.
I'd keep the permission with the asset, in a record the team can find, along with any limit on audience or time. A searchable archive is helpful, but availability shouldn't silently become a grant for every future project someone can imagine.
What to try this week
Before anyone feeds a file into an AI tool, have them write down the source, why they have access and the new use. Send that to the owner for a yes or no.