Questioning the Role of Moral AI as an Adviser within the Framework of Trustworthiness Ethics

Abstract

The main objective of this article is to demonstrate why despite the growing interest in justifying AI’s trustworthiness, one can argue for AI’s reliability. By analyzing why trustworthiness ethics in Nickel’s sense provides some wellgrounded hints for rethinking the rational, affective and normative accounts of trust in respect to AI, I examine some concerns about the trustworthiness of Savulescu and Maslen’s model of moral AI as an adviser. Specifically, I tackle one of its exemplifications regarding Klincewicz’s hypothetical scenario of John which is refracted through the lens of the HLEG’s fifth requirement of trustworthy artificial intelligence (TAI), namely, that of Diversity, non-discrimination and fairness.