What actually happens to your files in an AI-assisted workflow: retention, training use, NDA coverage, and the questions to ask any vendor.
- The risk is the tool, not the concept
- The questions that actually distinguish vendors
- How we handle it
The risk is the tool, not the concept
Pasting a confidential document into a consumer chatbot and running it through a controlled workflow are different activities that people describe with the same words. Most of the security anxiety around AI translation is really anxiety about the first one, which is justified.
Consumer AI tools generally reserve broad rights over submitted content, and their terms can change. A document handled that way has left your control in a way that is difficult to reverse.
The questions that actually distinguish vendors
Does content get used to train models? Where is it stored, and for how long? Does NDA coverage extend to the linguists and reviewers, or only to the company you contract with? What happens to the files at project end?
That third question separates providers more than any other. A vendor whose NDA covers only the corporate entity, while work is distributed to freelancers under no equivalent obligation, has a gap where it matters most.
How we handle it
NDA coverage extends to everyone in the delivery chain, including reviewers — not just the account contact. Angel Translation operates an ISO 9001:2015 certified quality system, and confidentiality handling is part of the documented workflow rather than a clause at the end of it.
If your project needs specific retention, storage or destruction terms, raise them at kickoff. They are much easier to build into a workflow than to retrofit onto one.