Google DeepMind warns of chaos as millions of AI agents go online

2026-06-11 · MIT Tech Review AI · Original

Google DeepMind is launching a major initiative to study the risks of millions of autonomous AI agents interacting online. Rohin Shah, leading AGI safety research, highlights the danger of these agents executing tasks without human supervision or following complex instructions from other networks, potentially leading to unpredictable outcomes.

Popular AI tools

CraiyonCraiyon

Free AI image generator (formerly DALL-E mini)

Leonardo.AILeonardo.AI

AI image generation platform for game assets and creative content

DALL-E 3DALL-E 3

OpenAI's latest AI image generator with precise text understanding

Pixlr AIPixlr AI

Online AI photo editor

Perplexity AIPerplexity AI

AI-powered search engine with conversational answers

CanvaCanva

Visual design platform with AI-powered tools