Narayanan: AI flooding of courts is the most under-recognized AI harm

random_walker · x · 2026-09-16

Arvind Narayanan argues "AI floods" — overwhelming communication channels with AI-generated slop — are the most under-recognized AI harm. Administrative agencies and trial courts are being flooded, driving adjudicators to use AI tools in response: "slop for slop, and the whole world goes blind." He explains why it's under-discussed: 50+ flood types register as separate problems, harms are diffuse, and they lack the vividness of agent-swarm hacks.

Related event: Princeton Professor Warns of Underestimated "AI Flood" Harm(2 posts)→

Original post →

More from Safety

Safety channel →