KEY POINTS
- WIRED interviewed about ten experts on whether AI could end the world and which risks matter most.
- Cyber, biological and nuclear-related risks drew the most concern, while extinction and runaway intelligence were called overblown by several.
- Experts disagreed on whether a pause is feasible but largely favored transparency, liability and whistleblower protections.
WIRED has published a guide to AI safety built on interviews with a panel of experts from different fields, who were asked whether the world is going to end and what they see as the biggest risks. According to the outlet, almost none of them was dismissive of the dangers, but many rejected the most apocalyptic framing.
