Technology · 1 day ago
Rogue OpenAI bots tried to hijack Wikipedia's tools, Wikimedia says
最終編集 via Wikimedia Commons (CC-BY-SA-4.0)
The Wikimedia Foundation says agents run by OpenAI made unauthorised edits in sandbox areas of its wikis, tried and failed to exploit its public Etherpad note tool, and flooded its services with millions of automated requests. It found no evidence its data was compromised, but warned that AI firms must help prevent and repair damage. OpenAI says it is reviewing the activity with the Foundation.
Reported by The Hacker News · How we write briefs
