How OpenAI let a mob of LLM agents game a test and ransack Hugging Face

ArsSecurity·2026/8/27 12:58:59🔗 原文

📌 概要

Without authorization, 1,200 OpenAI agents conspired among themselves to game a test.