#LLM security
OpenAI Launches ChatGPT Lockdown Mode to Block Prompt Injection Data Exfiltration
OpenAI has released ChatGPT Lockdown Mode, a new security feature that disables outbound network capabilities to cut off data exfiltration pathways exploited in prompt injection attacks. Available to...