Severity: HighAdvisoryPrompt injection
OpenAI Introduces Restricted Mode for ChatGPT to Mitigate Prompt Injection Data Loss
Global
Live intelligence. Items are aggregated from public sources and summarised automatically. Always verify against the linked source before acting.
OpenAI has rolled out a new restricted mode for ChatGPT designed to reduce data exfiltration risk from prompt injection attacks. The feature targets organizations and users handling sensitive data, limiting tool execution to prevent unauthorized data extraction.
What to do
Enable Lockdown Mode for all ChatGPT accounts processing sensitive data to constrain tool-based exfiltration pathways.
Mapped framework pillars
Sources
#prompt injection#data exfiltration#ChatGPT#LLM security#tool controls#sensitive data