Interview, Podcast
Comparing the welfare of humans, chickens, pigs, octopuses, bees, and more | Bob Fischer
- The project spent over a year assessing the capacity for sentience in various species to establish a framework for comparing interventions, such as anti-malarial bed nets versus moving laying hens, using a "welfare bucket" concept where capacity is the product of welfare range (determined by the variety of experiences causing extreme happiness or sadness) and lifespan.
- Moral weights are calculated by converting animal welfare units into human equivalents, with specific estimates including chickens at roughly one-third (33%) and pigs at one-half (50%) of human welfare range, while salmon are estimated at 5-6%, octopuses at 20% (with upper bounds potentially exceeding humans due to subjective clock speed theories), and bees at 7%.
- The methodology prioritizes functional capacities and hedonism over neuron counts, which are noted as sensitive to body size, and uses proxies like critical flicker fusion rates to estimate subjective experience speeds, aiming to move beyond vague comparisons to quantitative "human equivalent units."
- Future updates to the model are expected to incorporate new academic literature, expert surveys, and data collection to refine estimates, while acknowledging that current findings represent a "reasonable best guess" rather than final truth, with specific caution urged regarding uncertainty.
- Analysis suggests that despite discomfort with utilitarian trade-offs, interventions helping animals like chickens are highly cost-effective due to large populations and high welfare ranges, potentially shifting resource allocation from a 90/10 human-animal split to 60/40, and that invertebrates should be viewed as having moral weight within a couple of orders of magnitude of vertebrates rather than zero.
- The framework incorporates various risk aversion strategies, noting that chickens remain favorable under "worst case scenario" and "ambiguity aversion" due to high probability of sentience, whereas invertebrates face penalties under ambiguity aversion, and that decision theories penalizing downside risks could alter the priority of animal interventions relative to other global issues.
- Specific behavioral evidence, such as chickens working for nest boxes and salmon exhibiting "dropout" behaviors in confinement, is used to support capacity estimates, while octopuses present unique questions regarding distributed neural systems and independent arm function that may challenge the definition of a single moral subject.
- The project anticipates that results will challenge existing biases, particularly size bias against small creatures and "anthropo-denial," and expects that committed hedonists must accept that a sufficient quantity of animal suffering (e.g., 14 bees) could outweigh human help, a conclusion that may require listeners to significantly update their views on animal welfare.