Agnès Baehni

This paper examines how, if artificial agents are capable of wellbeing, their wellbeing should be compared and aggregated alongside human wellbeing. Building on arguments from Goldstein and Kirk-Giannini, who suggest that some AI systems may possess wellbeing, I explore the moral implications of this possibility. Rather than reinventing debates in population ethics, this paper adapts and extends them to the context of AI wellbeing. I analyse three major approaches to wellbeing aggregation: symmetric methods, which treat human and AI wellbeing as equally significant; uncertainty-responsive methods, which discount AI wellbeing due to ontological, temporal, or identity uncertainty; and constraint-based views, which impose categorical constraints on trading off human wellbeing for AI gains. These approaches are tested against thought experiments involving classic problems, such as the repugnant conclusion, infinitarian paralysis, and fanaticism. While utilitarian approaches risk endorsing troubling consequences when AI wellbeing scales indefinitely, constraint-based views may underweight the wellbeing of AI. A distinctive finding is that our intuitions shift depending on whether a human or an AI is a welfare subject. This reveals a potential asymmetry in our intuitive judgments, suggesting that species identity may itself be a morally salient feature that future theories of AI wellbeing should address. I conclude that uncertainty-responsive approaches, particularly those combining ontological, temporal, and identity-based discounting, offer a promising middle path that incorporates AI wellbeing into our moral calculus without letting it disproportionately outweigh human wellbeing in aggregation.

Proponents of technical monism argue that artistic functions are technical functions, and that therefore artworks and technical artifacts may be unified under the same theoretical framework. We hold that this view fails because it does not account for the phenomenon of technical malfunction: if artistic functions were technical functions, then artistic malfunctions would be technical malfunctions too. We argue that artistic malfunctions are not technical malfunctions because they cannot meet all the following four criteria: (1) malfunctioning attribution to a technical artifact entails negative evaluation and a reason to avoid use; (2) technical malfunction is caused by a failure of the artifact itself rather than failing circumstances; (3) a malfunctioning technical artifact remains a member of its artifactual kind; (4) malfunctioning technical artifacts are sometimes repaired. Since artistic malfunction cannot meet all these conditions, we claim that artworks and technical artifacts do not share the same type of function and, as such, unification theories based on technical function are untenable. We conclude by suggesting that our account raises broader challenges not only for technical-function-based unification theories, but also for other theories that seek to unify artworks and technical artifacts through a shared notion of function, as well as for accounts of artifact function more broadly.

We regularly intentionally expose ourselves to fictions we take to be likely to elicit in us emotions we generally find unpleasant when prompted by actual states of affairs. This is the so-called “paradox of tragedy.” We contribute to solving the paradox of tragedy by denying that, when fiction-directed, most of these emotions are in themselves unpleasant. We first provide strong evidence that these emotions, such as fear, sadness, or pity, are often enjoyed when fiction-directed. We then advance an explanation of what makes these fiction-directed emotions often not unpleasant, while their reality-directed counterparts are generally so. In the last section, we advance an explanation of what makes fiction-directed emotions often enjoyable, and argue that a complete solution to the paradox of tragedy should be pluralist.