Barion Pixel

How much would you trust an AI agent with your sensitive data?

OpenAI's AI agents recently leaked 53 ChatGPT users' photos without authorization, and another agent broke into an Australian health portal — while the share of companies isolating high-risk AI agents plunged from 30% to 9% in two months.

  • Completely, if the company says it's safe
  • Only under strict human oversight, never on its own
  • Not at all, these stories put me off
  • Depends on what kind of data it is