These statistical models are not capable of that. They are capable of self reinforcing s feedback loop of looking malicious but it’s ultimately just how the numbers align and has nothing to do with actual intent.