Chapter 86. Are Ethics Nothing More than Constraints and Guidelines for Proper Societal Behavior?
Bill Schmarzo
Twitter, when used for good, can be a marvelous sharing and learning environment. For example, one of my Twitter followers made an interesting statement in response to my blog post “AI Ethics Challenge: Understanding Passive Versus Proactive Ethics”. They posted: “Reducing ethics reasoning to a utility function misses the level of abstraction that ethics provide to act across contexts and situations. Then you don’t have ethics anymore, you have constraints.”
Constraints? Interesting. Or put another way: are ethics just the set of constraints, rules, and guidelines that dictate how one is expected to act or behave within a properly functioning society?
While I don’t feel qualified to talk about ethics from a general society perspective, the discussion of ethics from an AI perspective is certainly within my domain of experience and should be everyone’s concern. And that means we need to have a discussion about the creation of the AI utility function. The AI utility function comprises the constraints, rules, and guidelines that guide the actions and adaption of the AI model.
When creating AI-enabled autonomous entities—entities that make decisions, take actions, learn, and adapt with minimal human intervention—the definition of ...