Founders & Leaders
From p(doom) to Frameworks: How Experts Actually Assess Catastrophic AI Risk
The article critiques the use of "p(doom," a single probability estimate for AI existential risk, arguing it obscures underlying assumptions and expert disagreement. It proposes structured alternatives like the Belief/Plausibility model and the NIST AI Risk Management Framework for more effective risk assessment and governance.
Alejandro Mendoza·October 28, 2026