Thank you @alexeusgr
This was a very interesting read, as a tester (not only system testing but also in RL) I found this proposal very intriguing.
With the societal impact and your open problem statements in mind I could perhaps help define some of the solutions to those.
Guilt = generated when certain thresholds within the system parameters are triggered. Testing these thresholds and living on the edge, but also have an understanding of them is imperative to be able to work with the system and not against it. Thus enabling proper feedback.
Shame = suggestive information channel back to recipient.
The system algorithm should be set up with different levels, and also introduce different admonition tokens based on different guilt thresholds and not put them all in the same bucket.
The system must not be based off on religious premonitions of do’s & dont’s or other biases, but should rely on a global human norm that is lenient, but also more effective based on the different admonition tokens that are triggered.
The system should also include a contribution token, promoting effort and value contribution which overall could override the whole admonition tokens if the means supersedes and aligns with the end-goal.