GoKawiilF-Secure publishes taxonomy of 26 AI harms to consumers with 24 proposed defenses
F-Secure released an AI Harms Taxonomy, developed by researcher Dr. Megan Squire, that catalogs 26 distinct ways AI systems can harm consumers and traces how each harm travels through intermediary layers to reach a person. The framework spans five relationship types and 19 intermediary layers, and proposes 24 defenses aimed at specific points in that chain, covering risks such as voice and video impersonation and AI agents making damaging financial decisions.