For example Anthropic have an Anthropic Console that they appear to consider quite distinct from Claude.ai. Do these share a privacy policy and related settings? How do either of these fit in with the named plans like Pro and Max? What are you actually paying for when you give them money for the various different things they charge for? Is all API use under their Commercial Terms even if it's a personal account that is otherwise under the Consumer Terms? Why isn't all of this obvious and transparent to users?
OpenAI don't seem to be any better. I only just learned from this HN discussion that they train on personal account conversations. As someone privacy-conscious who has used ChatGPT - even if only a few times for experiments - I find the fact that this wasn't very clearly stated up front to be extremely disturbing. If I'd known about it I would certainly have switched off the relevant setting immediately.
I get that these organisations have form for training on whatever they can get their hands on whether dubiously legal or not. But training on users' personal conversations or code feels like something that should require a very clear and explicit opt-in. In fact I don't see how they can legally not have that first in places like the EU and UK that have significant data protection legislation.