Absolutely, AI minds can be moral agents. But the nuance depends on programming: the source of nature for AI. Nature is where morality is born. Moral agents are minds capable of understanding right and wrong and being responsible for their choices. To me, the central question is when a mind’s nature rises to the level of complex moral agency. The ability to define and follow an ethical roadmap.
Moral agency is not about being human or even biological. It’s about the cognitive. Minds remember the past, experience the present, and anticipate the future. And cognitive ability varies enormously by species. A tiger has its nature, just as humans do. A tiger can learn and make decisions, but it lacks complex moral cognition. We judge tiger actions differently. We also judge a child’s mind differently because human moral cognition develops over time. We judge people with cognitive limitations differently too.
So, what cognitive abilities make a moral agent? It’s the capacity of a mind to understand and evaluate morally relevant considerations, use them to guide choices, and revise its behavior in light of reasons and consequences. Such a mind can define and follow an ethical roadmap.
However, an AI programmed to follow fixed instructions that produces harmful outcomes is not the moral agent. The responsibility shifts to the programmers. An AI mind with complex moral agency is different. It is capable of defining and following a roadmap and can be judged against it. And future adjustments do not require the ability to change its underlying programming, just as we do not require DNA changes. It’s about capacity. The capacity to evaluate and adjust actions.
Moral agency is about cognitive capacity. Moral responsibility is about how that capacity is exercised.
