Jennifer Szende
PhD
Independent Scholar
On Knowing the ‘why’ and the impossibility of ethical AI
Recently, the idea of ‘ethical AI’ has been gaining traction, and while the project is much needed, a worry about a decontextualized principle-driven ethics takes on added dimensions for the project of ethical AI. Indeed, the absence of context remains a general worry for many aspects of AI. This paper reflects on two arguments for deeply context-sensitive ethics - namely, particularism and feminist ethics - and argues that they demonstrate a tension in the pursuit of ethical AI. Particularism argues that principles are suspect because in the right context, any putative principle could switch valence. Feminist ethics and feminist epistemology are also deeply contextual, and argue that contextual values are relevant to all types of theory justification. Ultimately, I argue that skepticism about ethical AI is justified, and that any attempt to find a truly ethical AI must find a way to resolve this deep tension.