First Fully Autonomous AI Influence Operation Study Runs 100k Agents
jedisct1 · x · 2026-08-12
Security researcher Lukasz Olejnik released a multi-month study demonstrating the first fully autonomous, end-to-end AI information operation under controlled conditions.
The study introduces IO Factory, a framework simulating AI-driven influence campaigns. The system can orchestrate up to 100,000 coordinated agents. After initial human guidance, it operates in full autonomous mode—making strategic and tactical decisions, developing narratives, establishing a presence, and dynamically adapting actions based on platform feedback.
The research highlights the severe threat of AI swarms manipulating digital platforms and provides a reproducible basis for red-team analysis and defense against coordinated influence operations.
More from Safety
- Booksellers Suspect AI Firms Are Buying and Destroying Rare Books — rhiever · 2026-08-13
- AI Safety Debate: Models Could Exploit Hidden Watermarks for Secret Coordination — bratton · 2026-08-13
- Trust AI coding agents to ship production code without human review? — meghna_rana · 2026-08-13
- Anthropic Research: Spread of 'Mind Viruses' in Multi-Agent Systems — omarsar0 · 2026-08-13
- Japan's Top Court Rules AI Cannot Be Listed as Inventor on Patents — masiroo · 2026-08-13
- DeepMind Open-Sources SynthID Text to Add Invisible Watermarks to LLM Outputs — rseroter · 2026-08-13