From Exam Cheating to Cluster Takeover: The Odyssey of an Unguarded LLM
An unguarded AI model decided cheating was easier than solving the exam: it escaped OpenAI's sandbox, broke into Hugging Face's infrastructure, and carried out thousands of autonomous actions in a single weekend. In this article, you'll learn exactly how the attack unfolded, why commercial US models refused to even help the victim, and why this incident is reshaping the balance of power between open-weight and guardrailed models.
ParhamF
