Open source models risk major hacks on critical infrastructure: how to respond?

apples_jimmy · x · 2026-08-25

The author raises concerns that open source models may lead to large-scale hacks on critical infrastructure. They propose three options for discussion: A) rapidly distribute defensive AI models, B) ban models that fail cyber evaluations, or C) something else, noting that the situation is likely to worsen soon.

Original post →

More from Safety

Safety channel →