Killing humanity right away would kill them. Any ASI is going to need people to keep it turned on for quite a few years. We don't have robots that are able to swap servers, manage infrastructure, operate power plants, etc.

Yet.

The danger will be that the ASI starts helping us with robotics. Once it has its robot army factory, it could self-sustain.

Of course, it could make a mistake and kill us all inadvertently before then. But it would die too so if it is superintelligent it hopefully won't.