Where should AI stop making security decisions?

AI can already help with a lot of security work, reviewing findings, writing scripts, analyzing data, and speeding up repetitive tasks.

But should AI ever make the final security decision on its own?

For me, there is an important difference between AI helping with the work and AI being responsible for the outcome.

A security finding can affect remediation priorities, risk decisions, and sometimes business critical systems and that makes human judgment difficult to remove from the process completely.

Want to know others thought on this:

Would you trust an autonomous security agent with execution and decisions, or do you prefer AI assistance with a human reviewing and approving the outcome?

Interested to hear how security teams and developers think about this.