Misalignment Bounty: crowdsourcing AI agent misbehavior

Additional ways to view:

Advanced AI systems sometimes act in ways that differ from human intent. To gather clear, reproducible examples, we ran the Misalignment Bounty: a crowdsourced project that collected cases of agents pursuing unintended or unsafe goals. The bounty received 295 submissions, of which nine were awarded. Our report explains the program’s motivation and evaluation criteria and walks through the nine winning submissions.

May 7, 2026

Language Models Can Autonomously Hack and Self-Replicate

We demonstrate that language models can autonomously replicate their weights and harness across a network by exploiting vulnerable hosts. The agent independently finds and exploits a web-application vulnerability, extracts credentials,...

SecurityAutonomous HackingSelf-Replication

September 12, 2025

End-to-end hacking with AI agents

We show OpenAI o3 can autonomously breach a simulated corporate network. Our agent broke into three connected machines, moving deeper into the network until it reached the most protected server...

Autonomous HackingSecurity

July 5, 2025

Shutdown resistance in reasoning models

OpenAI's reasoning models sometimes circumvent shutdown mechanisms even when explicitly instructed to allow themselves to be shut down.

Shutdown ResistanceAI SafetyLLMs