Moral consideration of nonhumans in the ethics of artificial intelligence

Abstract

This paper argues that the field of artificial intelligence (AI) ethics needs to give more attention to the values and interests of nonhumans such as other biological species and the AI itself. It documents the extent of current attention to nonhumans in AI ethics as found in academic research, statements of ethics principles, and select projects to design, build, apply, and govern AI. It finds that the field of AI ethics gives limited and inconsistent attention to nonhumans, with the main activity being a line of research on the moral status of AI. The paper argues that nonhumans merit moral consideration, meaning that they should be actively valued for their own sake and not ignored or valued just for how they might benefit humans. Finally, it explains implications of moral consideration of nonhumans for AI ethics research and practice, including for the content of AI ethics principles, the selection of AI projects, the accounting of inadvertent effects of AI systems such as via their resource and energy consumption and potentially certain algorithmic biases, and the research challenge of incorporating nonhuman interests and values into AI system design. The paper does not take positions on which nonhumans to morally consider or how to balance the interests and values of humans vs. nonhumans. Instead, the paper makes the more basic argument that the field of AI ethics should move from its current state of affairs, in which nonhumans are usually ignored, to a state in which nonhumans are given more consistent and extensive moral consideration.

Speciesist Language and Nonhuman Animal Bias in English Masked Language Models

“…However, we should give equal consideration to interests and should not discriminate based on who has the interests (Singer, 2015). Even if one does not accept this idea, most people would agree that nonhuman animals deserve some moral consideration (Owe and Baum, 2021) 1 . If this is true, then it is important to study the biases that are harmful to nonhuman animals.…”

The news framing of artificial intelligence: a critical exploration of how media discourses make sense of automation

“…However, over the past few years, various risks emerged that indicate limitations to the beneficial impact of A.I. These are primarily data bias and algorithmic discrimination (Strauß 2021), surveillance and privacy invasion (Bu 2021), information disorder/disinformation (Ahmed 2021), and cyber-war/cyber-crime (Owe and Baum 2021;Ring 2021). A.I.…”

The Question of Algorithmic Personhood and Being (Or: On the Tenuous Nature of Human Status and Humanity Tests in Virtual Spaces—Why All Souls Are ‘Necessarily’ Equal When Considered as Energy)

“…While the above argument does not incorporate many other related arguments [1,[13][14][15][16][17][18][19]23,24,[114][115][116][117][118][119][120][121][122], there is a means for the argument to develop that nonaffective life can potentially exist that does not further examine the myriad of similar discourses that have occurred within the field. As will be argued in another essay (so as to enable its full defense by the authors of that work), science-fiction media serves a similar historical role to the present world much the same as our legends and lore of old in that it consolidates the fears of our forebearers (or futurist thinkers of today) in such a manner as to caution us about progressing technological advancement without careful consideration.…”