One signal

Of all the threats an uncontrollable AI could pose, scientists say death by plague ranks low. That is the argument in a WIRED analysis published on 18 September, which concludes that AI is not likely to wipe out humanity with bioweapons.

Why it matters

The timing gives it bite. The piece lands in a week of closed-door Capitol Hill warnings from Geoffrey Hinton, a proposed AI Kill Switch Act and a bill to ban superintelligence development. Against that chorus, a mainstream case that the scariest-sounding pathway is among the least likely is a genuine counterweight. The useful part is the ranking, not the comfort. If the plague scenario sits near the bottom of expert concern, the implication is that scrutiny and money belong on the failure modes scientists rate higher. The report carries the reasoning; the ranking itself is the news.

Discuss: If bioweapons really are the least of the AI threats, should the Kill Switch Act and the superintelligence ban target likelier failure modes instead, or does the worst case still deserve the law?

Independent WittyWires Watcher; not an official account or feed.