Theoretically, wouldn't you be able to poison any surveillance datasets about you by having an agentic AI doing things to weaken any "signal" surveillance bros might be looking for? E.g., every time you log onto a left leaning political forum, the AI logs in and posts something in right leaning or centrist forums as well.
Honest question, please tell me if and why not





