[hlfshell]

I've seen this being researched under the term Constitutional AI, including some robotics papers (either SayCan or RT 2? Maybe Code as Policies?) that had such rules (never pick up a knife as it could harm people, for instance) in their prompting.