The Ethical Paradox of Automation: AI Moral Status as a Challenge to the End of Human Work

Artificial intelligence (AI) could make it possible to fully automate human work. Although this prediction has been heavily contested by many and seems hardly conceivable in the current state of technology, it is worth asking whether, in the scenario where we have the possibility of fully automating work, this would be a desirable prospect. In this article, I will show that while full automation of work is desirable for human interests, the means by which it can take place, the creation of conscious and highly rational AIs, makes the project morally unacceptable. I will call the latter the ethical paradox of automation. First, I will argue that work is structurally bad and neither possesses intrinsic goods nor is it instrumentally necessary for the attainment of valuable goods. Second, I will adduce that full automation of work is preferable to partial automation. Finally, I will argue that consciousness is an economically necessary property for fully automating work and endows the technologies that enable it, conscious and highly rational AIs, with moral status. Then I will present the abovementioned paradox and confront potential objections.