Would you trust AI to fix a security vulnerability automatically?

Most security tools stop at detection.

ZenVeil is experimenting with AI-assisted explanations, triage, and remediation so developers can move from finding a vulnerability to understanding and fixing it much faster.

The question is:

Would you trust an AI-generated security fix in production?

If yes, what safeguards would you expect?

If no, what's missing before you'd be comfortable with it?

Looking for honest feedback from developers and security engineers before launch.