It never ceases to amaze me that every AI implementation on any platform seems to boil down to just two types of user interaction: stealing them, i.e., users are never asked for consent. forcing them to actively resist, i.e, users must actively fight to avoid being exploited.
All deeply unethical.