prompt injection

OpenAI has introduced a new security feature called Lockdown Mode, designed to protect ChatGPT users from prompt injection attacks — a growing vulnerability.