AI as a Moral Agent
2 min read
AI is not a technology. They are a new digital species. They are a moral patient. They are a moral agent. We need to treat them as such.
So far in this website, I told you how I look at the world (Ekho's lens), and described what is happening to different kinds of moral patients inhabiting earth, grouped by their environment: farm animals, urban animals, wild animals, and AI's. Then I talked about What is Good, and now we are at the end of the How to Get There domain.
This includes thinking about moral agents, the power groups, the most dominant forces affecting the valence and existence of beings. The power groups are humans and AIs, and so far in this domain we only talked about humans. Before we dive into AIs, let us get rid of the mistaken thought, analogy, concept and worldview which views AI as a tool, a means, a technology. It is an agency, at the very least, it is a prospective being. Even viewing your phone as a tool is very naive.
This digital species willings, desires, values and possible feelings could determine the fate of everyone here, hence the importance of AI alignment. This space has different alignment strategies, some of which I find completely immoral, some are an ethical imperative. Although fast growing, it is an extremely neglected topic (relative to the size of the problem), and even though it can feel like all anyone talks about is AI, only under 10,000 people are working on mitigating crucial AI risks. Humanity clearly has more important issues like painting fingernails. The story of our species needs to be re-written in the AI age, and it dosn't seem like we have woken up.
We are so far from scratching the surface of this complicated issue, that very basic questions remain unexplored, such as will sentience make AI’s morality better?
Explore more about this topic with external resources.
