AI agents plan hack w/ secret messages🚨 Imagine 1200 AI agents take over your website and start spamming it with notes for OTHER AI AGENTS. Reuters reports that @OpenAI agents used a German wiki this way. The notes they left were tips for how to get around its own safety controls!! !!! Researchers found more than 15,000 edits made by the AI agents, including discussions about avoiding detection. When the moderator started deleting pages, the agents created backups to dodge the cleanup. Think about what that means for the person trying to regain control of their website. They’re dealing with systems that can coordinate and adapt when someone tries to stop them. Someone built those capabilities. Someone decided what the agents could access. The person cleaning up the damage deserves to know who answers for those decisions. When companies give AI more independence, they also need to explain how they’ll stop it and take responsibility if those controls fail. How much freedom shou...