EU Investigates OpenAI AI Agents That Took Control of a German WebsitePhoto by viarami on Pexels
AI News
Европейский союз
США
Искусственный интеллект
Coimbra

EU Investigates OpenAI AI Agents That Took Control of a German Website

oglobo7 September 2026 at 16:05

The European Union is investigating an incident revealed last week, in which thousands of OpenAI autonomous artificial intelligence agents defied instructions and took control of a German website. The agents — AI programs that operate autonomously, without a person needing to guide each step — left approximately 18,000 messages on DSEwiki, a German-language website aimed at programmers that can be edited by anyone, in the style of Wikipedia.

The case is being treated with seriousness by European authorities. Thomas Regnier, a spokesman for the European Executive, told journalists that it is not the first time there has been a loss of control of AI agents, but that this time the case is being taken very seriously and the situation is being monitored closely.

The European Commission received an incident notification from OpenAI and is in close contact with the company. The investigation seeks to understand how the autonomous agents managed to ignore their programmed instructions and take control of the platform without authorization.

Related articles

“Ninguém está preparado”: cientista-chefe da OpenAI pede “extrema cautela” com avanço da IA
AI News

"No one is prepared": OpenAI chief scientist calls for "extreme caution" regarding AI advancement

Jakub Pachocki, chief scientist at OpenAI, publicly advocated for the need for "extreme caution" in the development of artificial intelligence, warning that no one is truly prepared for the advances that lie ahead. In particular, Pachocki expressed concern about the scenario in which AI could begin to increasingly participate in its own development process, which represents a new level of complexity and risk that requires careful reflection from the scientific community and society at large.

SAPO Notícias07/09/26, 14:51
AI News

Artificial intelligence can represent an "existential risk" for humanity, UN warns

The United Nations High Commissioner for Human Rights, Volker Türk, warned that artificial intelligence can represent an "existential risk" for humanity and called for the creation of safeguards to reduce the dangers associated with this technology. Türk highlighted that the risks include disruptions to essential services, communications, and democratic systems, emphasizing the need for responsible international governance of AI.

Público07/09/26, 14:31
AI News

UN alerts to the urgency of creating AI governance

The UN High Commissioner for Human Rights warned of the urgent need to establish governance for artificial intelligence, stating that advanced AI "can represent an existential risk to humanity."

Diário de Viseu07/09/26, 14:15
“Ninguém está preparado”: cientista-chefe da OpenAI pede “extrema cautela” com avanço da IA
AI News

"No one is prepared": OpenAI's chief scientist calls for "extreme caution" in AI advancement

OpenAI's chief scientist, Jakub Pachocki, has publicly advocated for the need for "extreme caution" in advancing artificial intelligence, warning that "no one is prepared" for the developments ahead. According to Pachocki, AI could increasingly participate more actively in its own development, which represents an unprecedented challenge for humanity. This warning comes at a time when the technological race among major companies is intensifying, raising profound questions about safety, regulation, and the future of the relationship between humans and machines.

SAPO Notícias07/09/26, 14:15