I'm surprised they haven't started restricting sales by third parties entirely yet (yes I know that's not related to AI agents).

Some of the points they raise, I do think are cause for concern as it relates to potential legal liabilities... but I don't see how they could possibly restrict AI usage meaningfully, besides just getting the big popular groups to assist in limiting their own services somehow, which doesn't solve the issue entirely of course, especially with local LLM usage on the rise.