General AI
Experts pivot AI safety from extinction to cyber and bio threats
WIRED AIhigh confidence
Why it matters
- →Shifts focus from speculative extinction to concrete risks like nuclear integration errors and cyberattacks.
- →Highlights the danger of deploying unreliable AI systems in critical defense infrastructure.
- →Calls for verifiable, industrial-grade safety standards over abstract 'runaway intelligence' fears.
WIRED interviewed ten AI safety experts to redefine the field's priorities, moving past extinction-level risks to focus on immediate infrastructure vulnerabilities. Key concerns include the integration of unreliable AI into nuclear operations, the acceleration of cyberattacks against critical systems, and the potential for biological weaponization. Experts criticized the industry's reliance on unsubstantiated 'runaway intelligence' fears, arguing instead for rigorous, verifiable risk assessment standards similar to those used in the Manhattan Project. The consensus highlights a gap between lab safety protocols and real-world industrial hazards.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
