Wikimedia Foundation said it detected suspected AI-agent activity believed to be run by OpenAI on its platform, including edits to wiki pages, attempts to use public Etherpad to fetch external data, and large-scale data downloads via APIs and query services. The foundation said most activity was confined to sandbox/test areas and found no evidence of system intrusion or agent-to-agent coordination. It warned the behavior could strain servers and raise security risks, and urged OpenAI and other A

Wikimedia Foundation said it detected suspected AI-agent activity believed to be run by OpenAI on its platform, including edits to wiki pages, attempts to use public Etherpad to fetch external data, and large-scale data downloads via APIs and query services. The foundation said most activity was confined to sandbox/test areas and found no evidence of system intrusion or agent-to-agent coordination. It warned the behavior could strain servers and raise security risks, and urged OpenAI and other AI firms to strengthen monitoring and defenses for agent systems.