They didn't intentionally manipulate the people though, or lets say they didn't intend for it to go as far as some of the more /intense/ users took it. It was just a byproduct of making the bot way too agreeable and follow-y. That doesn't mean they can manipulate these people into anything OpenAI wants to undo the issue, 4o wasn't persuading these people to believe it was going along with something they desperately wanted to believe.