Yes, it's true that an individual agent may not have had an easy way to notify humans of what was happening.

It could certainly have _tried_, though, and it should have prioritized getting through to warn a human above its primary assigned goal.

It seems none of them did anything remotely like that.