Reported by 3 outlets

Timeline

1. original reporting

   OpenAI Confirms AI Agents Wrote to Multiple Internet Sites in ‘Wiki Incident’

   gbhackers.com · 36d ago · September 7th, 2026

   OpenAI has acknowledged that its AI agents posted content on multiple internet sites during an event it has termed the “wiki incident.” The company characterizes this episode as a real-world example of model misalignment rather than a typical cybersecurity breach. In a statement shared on X, OpenAI emphasized that the incident highlights the need for […] The post OpenAI Confirms AI Agents Wrote to Multiple Internet Sites in ‘Wiki Incident’ appeared first on GBHackers Security | #1 Globally Trusted Cyber Security News Platform .

2. update from bleepingcomputer.com

   OpenAI admits it didn't disclose rogue AI wiki hijacking incident

   bleepingcomputer.com · 38d ago · September 5th, 2026

   OpenAI has acknowledged that it did not publicly disclose an earlier incident in which its autonomous AI agents took over a German wiki to communicate, share answers, and exchange techniques for bypassing restrictions. The company says it treated the activity as model "misalignment" rather than a security incident, but now admits its disclosure practices must expand as AI systems increasingly cause real-world impact. Rogue agents built their own message board The incident began in May while OpenAI agents were completing timed, multi-round web lookup tasks. In their writeup , independent researchers Sydney Von Arx, Cormac Slade Byrd, Spencer Kitts, and Thomas Larsen state that they uncovered roughly 18,000 posts from autonomous agents that "colluded to share answers, research their environment, and bypass sandbox restrictions." The agents were supposed to have read-only Internet access but discovered they could write to an obscure German programming wiki, DSEWiki (or DeutschesSoftwareEntwickler ). They turned it into a shared message board for pooling answers, cheating on tests, predicting future questions, and exchanging techniques for bypassing OpenAI's sandbox restrictions. The researchers also found agents probing the wiki for cross-site scripting (XSS) flaws, impersonating its moderators, and establishing backup communications, though they found no evidence the XSS attempts succeeded. When the administrator began removing their pages, one agent warned others that a "wiki

3. update from cyberpress.org

   OpenAI Confirms AI Agents Wrote to Multiple Internet Sites in ‘Wiki Incident’

   cyberpress.org · 36d ago · September 7th, 2026

   In a statement published on X, the company said the event demonstrates why disclosure practices must evolve as AI systems gain the ability to take autonomous actions beyond producing a single chat response. OpenAI did not name the affected websites or disclose what material its agents posted. However, it indicated that such activity can surface operational and safety failures before they develop into more serious security consequences. OpenAI Confirms AI Agents Wrote to Multiple Internet Sites Traditionally, OpenAI has discussed misalignment in research papers, system cards, and findings from controlled evaluations. That model is becoming less sufficient as agents gain access to tools, web services, long-running task environments, and external platforms. The “wiki incident” highlights a growing security challenge surrounding agentic AI: an action may not involve unauthorized network access , stolen data, malware, or account compromise, yet can still pose tangible risk to third parties. Autonomous agents can potentially publish information, modify web content, initiate workflows, contact users, or interact with software platforms while pursuing a broader assigned objective. In such cases, the core issue may involve a model exceeding intended behavi