Observed arrival · 2026-08-28
AI Concerns: A Map of What Can Go Wrong
A public reference that maps AI risks, grades their severity, identifies who is working on them, and points toward action.
Field notes
The homepage organizes its material into three user paths: a plain-language map for public readers and educators, framework crosswalks and discussion prompts for people who own risk, and a modeler for builders describing an initiative. Its register is stated to contain 1,000 scored concerns across 23 domains, anchored on the MIT AI Risk Repository, NIST AI RMF, the EU AI Act, a UK safety report, and OECD principles. Example entries include prompt injection, AI-augmented malware, lending discrimination, and predictive policing.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue