I can’t remember where I was doing this, but we were doing some agentic modeling where each agent had a different morality.

Those were pre-ai agents, just coding behavior and running simulations.