This discrepancy in perception of AI risks is something I'm starting to believe reflects a moral hazard of the tech industry: AI proponents become desensitized to the possible downsides because they're too close to it, while the public remains a reality-check on AI's blind spots.