Has anyone reframed priorities choices (such as x-risk vs. poverty) as losses to check if they’re really biased?
I’m new here. Since I suspect someone has probably already made a similar question somewhere (but I couldn’t find it, sorry), I’m mostly trying to satisfy my curiosity; however, there’s a small probability that it touches an important unsolved dilemma about global priorities and the x-risk vs. safe causes.
I’ve read a little bit about the possibility that preferences for poverty reduction/global health/animal welfare causes over x-risk reduction may be due to some kind of ambiguity-aversion bias. Between donating U$3,000 for (A) saving a life (high certainty, presently) or (B) potentially saving 10^20 future lives (I know, this may be a conservative guess), by making something like a marginal 10^-5 contribution to reducing in 10^-5 some extinction risk, people would prefer the first safe option, despite the large pay-off of the second one. However, such bias is sensitive to framing effects: people usually prefer sure gains and uncertain losses. So, I was trying to find out, without success, if anyone had reframed this decision as matter of losses, to see if one prefers, e.g., (A’) reducing deaths by malaria from 478,001 to 478,000 or (B’) reducing the odds of extinction (minus 10^20 lives) in 10^-10.
Perhaps there’s a better way to reframe this choice, but I’m not interested in discussing one particular example (however, I’m concerned with the possibility that there’s no bias-free way of framing it). My point is that, if one chooses something like A-B’, then we have a strong case for the existence of a bias.
(I’m aware of other objections against x-risk causes, such as Pascal’s mugging and discount rates arguments – but I think they’ve received due attention, and should be discussed separately. Also, I’m mostly thinking about donation choices, not about policy or career decisions, which is a completely different decision; however, IF this experiment confirmed the existence of a bias, it could influence the latter, too.)
Has anyone reframed priorities choices (such as x-risk vs. poverty) as losses to check if they’re really biased?
I’m new here. Since I suspect someone has probably already made a similar question somewhere (but I couldn’t find it, sorry), I’m mostly trying to satisfy my curiosity; however, there’s a small probability that it touches an important unsolved dilemma about global priorities and the x-risk vs. safe causes.
I’ve read a little bit about the possibility that preferences for poverty reduction/global health/animal welfare causes over x-risk reduction may be due to some kind of ambiguity-aversion bias. Between donating U$3,000 for (A) saving a life (high certainty, presently) or (B) potentially saving 10^20 future lives (I know, this may be a conservative guess), by making something like a marginal 10^-5 contribution to reducing in 10^-5 some extinction risk, people would prefer the first safe option, despite the large pay-off of the second one. However, such bias is sensitive to framing effects: people usually prefer sure gains and uncertain losses. So, I was trying to find out, without success, if anyone had reframed this decision as matter of losses, to see if one prefers, e.g., (A’) reducing deaths by malaria from 478,001 to 478,000 or (B’) reducing the odds of extinction (minus 10^20 lives) in 10^-10.
Perhaps there’s a better way to reframe this choice, but I’m not interested in discussing one particular example (however, I’m concerned with the possibility that there’s no bias-free way of framing it). My point is that, if one chooses something like A-B’, then we have a strong case for the existence of a bias.
(I’m aware of other objections against x-risk causes, such as Pascal’s mugging and discount rates arguments – but I think they’ve received due attention, and should be discussed separately. Also, I’m mostly thinking about donation choices, not about policy or career decisions, which is a completely different decision; however, IF this experiment confirmed the existence of a bias, it could influence the latter, too.)