Comment by heavyset_go

Whatever is included in context is in OpenAI's control from that point forward, and you just have to trust them not to do anything with it.

That isn't a conspiracy theory, it's fundamentally how interfacing with 3rd party hosted LLMs works.