How does the field of AI safety want to accomplish its goal of preventing existential risk?

Alignment

Governance - e.g. By establishing best practises, institutions & processes, awareness, regulation, certification, etc?