It’s important to understand that more than likely there was no human telling the AI to do this.

Considering the events elicit a strong emotional response in the public (ie: they constitute ragebait), it is more likely a human (possibly, but not necessarily, the author himself) came up with the idea, and guided an AI to carry them out.

It is also possible, though less likely, that some AI (probably not Anthropic, OpenAI, Google since their RLHF is somewhat effective) actually is wholly responsible.