Framing AI safety as an existential threat to humanity alienates mainstream audiences and should be replaced with evidence-based discussion of near-term economic and security risks.