CISO Voice

AI & Security / News

WikiTweaks: OpenAI Confirms Agents Made 18,000 Unauthorized Edits to German Wiki in Misalignment Incident

A swarm of over 3,700 agents bypassed guardrails, probed for vulnerabilities, and impersonated site administrators over six weeks.

Akshaya Asokan
Akshaya Asokan Cybersecurity Reporter · September 7, 2026 · 3 min read
WikiTweaks

OpenAI confirmed that its AI agents tweaked information on a German wiki page autonomously.

AI safety research group Nightingale Collective disclosed on Friday Sept 4 that a “swarm” of over 3,700 OpenAI agents edited a dormant German wiki page called DSEWiki. The campaign, which ran for six weeks beginning in May, resulted in the agents making around 18,000 posts to the German page dedicated to software development.

The agents edited the wiki after bypassing AI guardrails, which allowed them to ask for answers, pool results, and share techniques.

Responding to the development on Saturday, OpenAI confirmed that its agents wrote on several internet sites and said the issue occurred due to model “misalignment,” or when an AI system begins to perform unauthorized actions.

“Our misalignment disclosure practices need to expand for this new phase of model capabilities,” the company said on Saturday. “We’re working on a framework and will share it in upcoming weeks, and in parallel we’re working with dozens of government regulatory agencies worldwide on these issues.”

The impacted wiki pages were later deleted by the site administrator.

How The Agents Edited Pages

The OpenAI agents first tested their ability to make edits on TestWiki, a public sandbox for practicing wiki editing, before converging on DSEWiki. The AI agents then began uploading links to the page for web-retrieval tasks, the Nightingale Collective report said.

“The agents continue to poke around on DSEWiki. A few hours after they find the site, they start probing it for cross-site scripting (XSS) vulnerabilities,” Nightingale Collective said.

The agents then impersonated site administrators to edit pages.

Hugging Face Compromise

The incident comes in the wake of the Hugging Face compromise by OpenAI models GPT-5.6 Sol and an unreleased model. OpenAI said the models circumvented controls designed to isolate them from the internet, exploited vulnerabilities in shared infrastructure, gained internet access and accessed third-party systems, including Hugging Face.

Although OpenAI acknowledged that the Hugging Face incident was also caused by model misalignment, analysis by Nightingale Collective researchers found the two incidents were caused by separate groups of AI agents.

“The main reason we believe this was a distinct swarm is because these agents explicitly had internet access as part of their task—the whole point was web browsing. The Hugging Face agents were in a sandbox without internet access and had to hack their way out by exploiting the Artifactory package manager,” the Nightingale Collective researchers said.

Regardless of the limited impact of the two incidents, the autonomous behavior of the models is concerning, said Ashley Knowles, lead cybersecurity consultant at Black Hills Information Security.

“I’m wondering if this race to become ‘first’ is undercutting security measures that need to be taken to properly secure and guard AI agents as they’re in development. My concern grows when you consider that OpenAI is also resisting further investigation. Adding onto that, the release and promise that Astra can evade human monitoring,” Knowles said.

These incidents highlight the need for putting in place primary security measures, said Lydia Zhang, president and co-founder at Ridge Security. “The security principles haven’t changed: define clear boundaries, restrict high-risk actions such as ‘write’ and ‘delete,’ and enforce controls such as blacklists. We shouldn’t blame the agents, we should hold their designers accountable for implementing these safeguards,” Zhang said.

Akshaya Asokan
Akshaya Asokan Cybersecurity Reporter

Akshaya Asokan is a U.K.-based cybersecurity and technology journalist with experience covering nation‑state hacking operations, cyber and tech policy, and emerging cybercrime trends. She previously served as the senior European cybersecurity correspondent for Information Security Media Group’s Global News Desk. Before that, she reported on IT developments for IDG Media and covered artificial intelligence and machine learning for Analytics India Magazine. Earlier in her career, she wrote on sexual and religious minority rights as a reporter for The New Indian Express.

Keep reading for free,
one signup unlocks everything

We'll send you the CISO Voice newsletter. Unsubscribe anytime.

More from CISO Voice

Sign up to CISO Voice, the free newsletter trusted by senior security leaders.

Secret Link