Comment by hbrn

If AI does something good, you take responsibility. If it does something bad - "the agent messed up". /s