Gary O'Brien

Recent developments in large language models and image generation software raise the possibility that AI systems might one day replace humans in some of the intrinsically valuable work through which humans find meaning in their lives – work like scientific and philosophical research and the creation of art. If AIs can do this work more efficiently than humans, this might make human performance of these activities pointless. This represents a threat to human wellbeing which is distinct from, and harder to solve, than the automation of merely instrumentally valuable activities. In this paper I outline the problem, assess its seriousness, and investigate possible solutions. I argue that AI could reduce our incentives to perform such work, and this might result in a great deskilling of humanity. Furthermore, even if humans continue to do such work, the mere existence of AI systems would undermine its meaning and value. I critique Danaher’s (2019a) and Suits’ (1978) arguments that we should embrace the total automation of work and retreat to a ‘utopia of games’. Instead, I argue that the threat to meaning and value posed by AI gives us a prima facie reason to slow down its development.

Totalism states that one population is better than another iff it has higher total welfare. One counterintuitive consequence is the Repugnant Conclusion (RC). Totalism also entails that a very large population of animals with lives barely worth living is better than a smaller population of happier humans. Furthermore, the strategies that have been used to avoid the troubling normative implications of the RC do not work in the animal case, so we may have reason to bring about such a population. I introduce the notion of ‘Efficiency of Welfare Production’ – that animals of different species vary in the efficiency with which they convert resources into welfare. If we want to maximize total welfare, without any speciesist bias, we should identify which species is most efficient and try to maximize the population of that species. This has counterintuitive implications whether we accept hedonism or a more sophisticated theory of welfare.

Longtermism is the view that positively influencing the long-term future is one of the key moral priorities of our time. Longtermists generally focus on humans, and neglect animals. This is a mistake. In this paper I will show that the basic argument for longtermism applies to animals at least as well as it does to humans, and that the reasons longtermists have given for ignoring animals do not withstand scrutiny. Because of their numbers, their capacity for suffering, and our ability to influence their futures, animals ought to be a central concern of longtermists. Furthermore, I will suggest that longtermism is a fruitful framework for thinking about the wellbeing of animals, as it helps us to identify actions we can take now that have a reasonable chance of improving the wellbeing of animals over the very long term.

There are different meanings associated with consequentialism and teleology. This causes confusion, and sometimes results in discussions based on misunderstandings rather than on substantial disagreements. To clarify this, we created a survey on the definitions of ‘consequentialism’ and ‘teleology’, which we sent to specialists in consequentialism. We broke down the different meanings of consequentialism and teleology into four component parts: Outcome-Dependence, Value-Dependence, Maximization, and Agent-Neutrality. Combining these components in different ways we distinguished six definitions, all of which are represented in the philosophical literature. We asked the respondents which definition is best for consequentialism and for teleology. The most popular definition of consequentialism was the one which accepted value-dependence, but not maximization and agent-neutrality. We therefore recommend the use of this meaning to avoid misunderstandings. The results for teleology were more problematic, with several respondents claiming they never use the term, or indicating that it is confusing.