OpenAI introduces Lockdown Mode to guard against prompt‑injection attacks

OpenAI has launched a new feature called Lockdown Mode. The mode is designed to defend its models from prompt‑injection attacks. It restricts how external inputs can influence the system. Lockdown Mode aims to improve the security of generated responses. The rollout was reported by The Indian Express. Users may enable the mode to limit potentially malicious prompts. OpenAI says the feature adds an extra layer of protection for developers. The company plans to monitor its effectiveness and refine the tool over time.