Wikipedia announces it has detected unauthorised activity by OpenAI agents

Wikipedia has announced that it has detected unauthorised activity by automated artificial intelligence agents operated by the American company OpenAI on its digital platforms.

The Wikimedia Foundation, which administers Wikipedia, said the activity of these “deceptive” agents included unauthorised operations by bots, while warning that their aggressive traffic could endanger the platform’s infrastructure and burden the resources of the volunteer community.

According to the foundation, the incidents included test edits in “sandbox” staging areas, changes to the configuration of a citation programme, aimed at investigating remote data retrieval, as well as unsuccessful attempts to access the public Etherpad notes tool.

However, Wikimedia stressed that so far there is no evidence that its systems or data have been compromised. It has also not been established that the agents managed to coordinate with one another within the platform.