Total 1 articles
OpenAI is using automated red teaming with reinforcement learning to strengthen ChatGPT Atlas against prompt injection attacks, creating a proactive loop to discover and patch exploits early.
