The future can be enormously valuable if it contains a large population of sentient beings. Multiple authors have noted that, to improve the value of the long-term future, we should reduce the risk of human extinction. This would increase the expected size of the future population, which is good if one expects the average moral value of a life to be positive. However, a future without human extinction is not necessarily more valuable than a future in which humanity goes extinct. To assess the value of reducing human extinction risk, I first look at the expected moral value of the long-term future. Moral uncertainty between Totalism and Asymmetric Views makes it difficult to assess the expected moral value of a single possible future. However, even if the future would be worse than extinction, reducing the risk of human extinction risk could still be positive; reducing the sources of human extinction risk will also reduce the risk of global catastrophe. I argue that we should expect global catastrophes to have a negative effect on the value of the long-term future if they do not lead to extinction. If global catastrophes are unlikely to lead to extinction, this would be a reason in favour of reducing the sources of extinction risk. In conclusion, the expected moral value of human extinction risk reduction depends on one’s moral uncertainty between Totalism and Asymmetric Views and on the likelihood of global catastrophes to lead to extinction.
Key Words: longtermism; moral uncertainty; extinction risk; existential risk; global catastrophic risk; expected moral value
This thesis was both amazingly exciting and a huge pain to write. I think the strong part is that, at the time, it was making a point that wasn’t made in the literature (i.e. that the expected value of the long-term future is not necessarily positive). It is written clearly, and the narrative structure is clear. It also includes a clear formal setup. Weak parts include that I had to make the setup very narrow for this topic, which makes it much less useful. In addition, I think the empirical arguments are actually not so strong, even though they decide the conclusion.
I changed my mind about a number of things, but also consolidated some pre-existing beliefs. I updated that extinction risk reduction depends less on the value of the future, and more on how the action itself affects the long-term trajectory. In short, hazards that can cause extinction can also cause a lot of suffering or generally hurt the long-term trajectory of sentience. I learned a lot more about suffering-focused views, but remain skeptical about them. Practically, I learned that choosing an original topic can be difficult and lead to problems down the line, so it’s good to be prepared for that. Another practical lesson is to always check with your audience or reviewers to see whether you are meeting their values and expectations.
I think the topic is important because many laypeople seem to, on first reflection, not care about human extinction because they believe the future looks negative in expectation. Although I tentatively disagree, I think the general point that ‘we should consider what the future would look like’ is a valid point.