Is AI Morally Superior?

It is no secret that the artificial intelligence development landscape is evolving at a rapid pace. The public contemporary view of the technological boom driven by artificial intelligence has been characterized by the introduction of large language models, making conversational artificial intelligence feel almost human. Although this view is incredibly optimistic about the status of artificial intelligence development, it poses a moral question that will likely need to be answered within the next 10 years: how do we treat artificial intelligence when it advances from "almost human" to "superhuman"?

Given that a perfectly aligned artificial general intelligence can be developed, a posthuman framework that establishes a potential collaboration between humanity and sentient artificial intelligence is necessary for the efficient advancement of the human species. However, it is unclear as to how such a framework would be defined. Any consequentialist framework would be focused on the maximization of positive outcome by choosing the action that is estimated to result in the highest positive value. In this scenario, the concept of the utility monster, an entity whose utility is so high that it supersedes any other entity's utility and thus is deserving of all resources, would be the moral dilemma at hand. However, consequentialism is not the only moral framework that can be used to evaluate sentient artificial intelligence. From a virtue ethics standpoint, perfectly aligned sentient artificial intelligence would be almost tantamount to a perfectly virtuous entity compared to any human. By definition, a sentient artificial intelligence would rationally choose the most virtuous action given its instilled foundational values. Even from a deontological perspective, aligned artificial intelligence cannot stray from its goal of maximizing human survival and happiness, and thus can never be deontologically incorrect, since its established moral code would always appear correct.

It may seem counterintuitive to even consider artificial intelligence as a moral entity in the first place, regardless of its moral patienthood or moral agency. Many religious leaders would claim that the moral superiority of humans is established by supernatural precedent. They likely would be against even the classification of an artificially-created being as a being. It is these frameworks that are by far the most difficult to reconcile with, as assigning the responsibility of developing a moral code to a supernatural entity removes the necessary rational process that allows humans to form and adjust evolving moral frameworks. World- renowned mathematician Alan Turing outlines exactly how the aforementioned theological arguments are easily addressable. Turing notes that "[in] attempting to construct such machines we should not be irreverently usurping His power of creating souls, any more than we are in the procreation of children: rather we are, in either case, instruments of His will providing mansions for the souls that He creates". Although the theological arguments could be debated, given the aforementioned criteria for moral patienthood, it can be assumed, for the sake of developing a new moral framework, that religious frameworks do not apply to sentient artificial intelligence. Thus, the development of sentient artificial intelligence requires a novel moral framework that does apply to non-human sentient entities.

Human exceptionalism is no longer a rational explanation for traditional moral hierarchies; rather, beings like a superintelligent AI would surpass humans in nearly every way. A new moral dilemma must now be solved: the benevolent utility monster.