I’ve heard people argue that it’s possible global health and development could actually be negative utility due to the consequence of increased meat consumption and also factory farming. It feels like EAs who value animal suffering at all must provide very clear reasons why its okay to support saving the lives of meat-eaters and also developing the third-world given the “meat eater problem.”
Even stepping away from utilitarianism, it seems more wrong to actively save someone who is very likely going to commit a moral atrocity (if you believe eating meat is a moral atrocity).
How do you deal with the meat-eater problem? I find the problem very compelling and I do not know myself how to deal with it.
Since the discussion on this thread, I’ve had the view that the meat-eater problem is dwarfed by the cause prioritisation problem, in the sense that if you give money to a global health and development charity, overwhelmingly the biggest harm to animals is that you didn’t give that money to animal welfare charities: the actual negative effect of your donation is likely very small by comparison.
(There’s obviously an act-omission difference here, but I don’t personally find that an important difference.)
Im worried that the chunk of EA that is concerned with effective human nearterm charities are all at risk of being net negative
Hi Sammy,
I think describing donations to global health and development interventions as net negative because animal welfare interventions are more cost-effective is misleading:
The default counterfactual for this sort of comparisons is simply not donating anything, i.e. keeping more money for personal consumption.
I believe cost-effective global health and development interventions, such as those of GiveWell’s top charities, are more cost-effective than the marginal personal consumption of donors.
Just to be clear, I also see value in Ben’s point. As I had answered below:
Sorry I meant compared to doing nothing. I’m mainly concerned about specifically the consequence of increased meat consumption and also factory farming from GH&D.
Ah, sorry for misinterpreting too. I thought your “net negative” was connected to Ben’s point.
Being net negative or positive and how much just depends on the values of whoever does this assessment. So I don’t think such statements are useful. These EAs may be net negative given your values. Probably much less so given their values.
I don’t think it is useful or helpful to speak in general terms about how positive/negative the (expected) value of something is. There is no universal way to value stuff.
It is a pretty uncomfortable problem and not one that I have been able to reconcile very well. One way around it is steering people to support global health/dev orgs that help people while not increasing meat consumption. An example is
FemaleFamily Empowerment Media, which improves openness and access to contraceptives in Nigeria. Another examples is the Beans is How Coalition, which aims to double worldwide bean consumption for the purpose of reducing hunger and increasing sustainability.Tangentially, this conversation illustrates how (if person-affecting views are false), the sign of Family Empowerment Media (FEM) is the opposite of AMF and other life-saving charities. FEM prevents human lives and AMF saves lives, and they have the opposite downstream effects on human lived experience, farmed animal welfare, and so on.
Therefore, I would not suggest anyone split their donations between life-preventing charities like FEM and lifesaving charities like AMF, because their effects will offset each other. People who are sympathetic to FEM (as opposed to AMF) because of farmed animal effects should probably just donate to animal welfare charities which I would expect to help animals even more.
The sign is only opposite through this particular generic population increase/decrease channel though.
AMF also has impacts on quality of life and maybe human capital I suspect.
FEM may have positive impacts on earnings, on women’s rights, and on the composition of who has children and when (ie towards people who are ready and willing to have them).
I agree with that caveat! Though I suspect that the downstream effects of the population increase/decrease channel dominate, especially for animal welfare.
Cool org I’ve not heard of, thanks!!
EDIT: Thanks Richard, slightly silly question in retrospect!
Thanks Constance, how is FEM better (I think Family not Female :D) from this perspective than any other “saving life” org, like AMF?
I think the idea is to reduce the future population of meat-eaters by encouraging contraceptive use, so kind of the opposite (in terms of total population) of saving lives.
(I have to say, the idea that we should positively prefer future people to not exist sounds pretty uncomfortable to me, and certainly less appealing than supporting people in making whatever reproductive decisions they personally prefer, which would include both contraceptive and fertility/child support.)
Your writings on this subject often emphasize an extremely high regard for the value of people making their own reproductive decisions, even when the weights are (as in this case) a human’s life and an enormous amount of farmed animal suffering.
When would the other stakes be sufficiently large for you to endorse preventing someone from making their own reproductive decision?
For example, let’s say Hitler’s mother could have been forced to have an abortion, preventing Hitler’s birth. Would you say that’s a tradeoff worth making, with regret?
Or let’s say we know Alice’s son Bob, were he to be born, will save 1 billion lives by preventing a nuclear war, and Alice currently intends to abort Bob. Would you say forcing Alice to carry Bob to term would be a tradeoff worth making, with regret about the forced birth?
The reason why I ask is because my intuition is that while reproductive autonomy is very important, it seems to me that there are always ways to up the stakes such that it can be the right thing to compromise on that principle, with regrets. I feel like there’s something I’m missing in my understanding of your view which has caused us historically to talk past each other.
If you can stipulate (e.g. in a thought experiment) that the consequences of coercion are overall for the best, then I favor it in that case. I just have a very strong practical presumption (see: principled proceduralism) that liberal options tend to have higher expected value in real life, once all our uncertainty (and fallibility) is fully taken into account.
Maybe also worth noting (per my other comment in this thread) that I’m optimistic about the long-term value of humanity and human innovation. So, putting autonomy considerations aside, if I could either encourage people to have more kids or fewer, I think more is better (despite the short-term costs to animal welfare).
Thanks of course! That’s actually quite obvious in retrospect, not sure how I missed that on first pass.
There would also be a counter argument that reducing family size is strongly associated with rapid development and with it in turn mass deployment of factory farming. One American probably eats 10-30x the meat of the Nigerians FEM serves. Its tricky....
You would still have to deal with the increase in factory farming and per capita meat consumption that comes with societal development.
This doesn’t answer the core question, but in many places like UgAnda the lives we are saving (people who live in the village) aren’t eating any factory farmed meat. The little meat they do eat is from animals reared locally which I think have net positive lives.
I have recently done a bit of research on the intensification of animal agriculture in Africa. I have a few comments to make in response to yours.
I am very confident that people in poor countries like Uganda eat way less animal products than the global average. But I am not sure that they all don’t eat factory farmed animal products. I think I have quite a high level of belief that your claim about the meat consumption patterns of the people in the areas in Uganda you work in. But I don’t think we should generalise to: “All people in very poor countries don’t eat factory farmed meat”.
I think a very important fact we should recognize is that factory farming clearly exist and is booming and intensifying quickly in Africa, including Uganda, or even poorer countries such as Burundi and South Sudan. This means that the meat-eating problem (convinced by JWS’s comment that we should change the wording, even though I don’t agree about all the things said in the comment), if it is a problem at all, is going to get worse in Africa and other parts in the world with many people in extreme poverty.
A very important note needs to be introduced here: I think we one species of farmed animals we should focus a lot on is the chicken (and also fish farming, maybe in 3-5 years time). Some facts about chicken farming in Africa:
It’s one of the cheapest type of meats poor people can afford, in a lot of regions.
Many aid providing foundations/agencies/charities are interested in, if not executing, using chicken farming as a poverty aid intervention. (1, 2, 3, 4)
While it’s hard to find a single broiler chicken farm in North America and Europe raising broiler chickesn in cages (with the exception of Russia and Ukraine—no intention to create drama), there seems to be currently a large wave of new broiler farms emerging in Africa that are going for caged broiler systems
For instance, chicken farming equipment producers in China are very actively trying to sell broiler cages to Africa, and according to my research, 10 broiler cage producers from China (out of 18 I found) are trying to do that.
(Sorry, I can’t share the links or details of my research on this topic here, as I worry about potential info hazards.)
The emergence of new technologies (such as modern caged broiler systems, digital gadgets, and eventually AI) and the popularization of both newer and older technologies (such as vaccines, drugs, feed formulas, and caged layer systems) will drive costs down further.
I think the rise of intensified, caged system raised chickens (both layers and broilers) in Africa (also some countries in Asia and Latin America) should alert and worry us that the “meat eating problem”, if it does not pose a huge problem now, could become much more severe in the future because of the rising per capita consumption of animal products coming from horrific systems. While there might be a lot of strategic/signalling/philosophical issues thinking and calling life-saving or poverty alleviating interventions as “saving meaters/people who harm animals”. I think we should definitely oppose to making things worse for animals in order to lift people out of poverty.
A less important response to your another point: I think it’s very unclear whether farmed animals raised in free range condittions in poor countries live net-positive lives. Firstly, many deadly diseases are very common among free-range (and intensive) chicken farming, such as Newcastle Disease. And awareness to use vaccines to reduce such diseases is nowhere near widespread. Secondly, debeaking of chickens is very common (including in Uganda), and presumably mostly without anesthesia/pain relieve. Thirdly, it seems very common to transport chickens like this (and in some cases kept tied like this even after arriving at the market, until they are sold) in Africa and poorer parts of Asia. Fourthly, I saw some nasty slaughters—let’s say they are at least as nasty as most chicken slaughters in the world. Finally, we have to consider that chickens raised for meat don’t live for many days in their life to compensate for these pretty intense suffering. IMO it’s more likely than not that most chickens raised in free range conditions in poor countries live net-negative lives.
(I weak upvoted your comment and chose “disagree”, even though I don’t 100% disagree with you.)
Thanks Fai—I was just making a small comment, to point out that most rural Ugandans, and probably Sub-saharan Africans eat either little or non factory farmed meat.
To your comment “But I am not sure that they all don’t eat factory farmed animal products.” For sure, plenty of people here eat a LOT of barn/factory farmed meat (mainly chickens see below), those in cities—but not the 70% of people who live in the village and barely ever buy “meat” at all.
I agree its unclear whether free range farm animals live net positive lives—I’m maybe 70% sure they do. There’s no de-beaking in the village here—that starts happening when animal farming is commercial. For sure transport is often nasty, Slaughter is often especially horrible but in my wee opinion that’s not nearly enough to negate the rest of their lives doing what animals do without huge constraint. We can probably agree to disagree on the whole-life positivity vs. negativity thing. Yes chicken lives are short, and many die early to disease (which can but doesn’t always involve a lot of suffering). Most village chickens here though live between 4 and 9 months—it takes 6 months-ish before they reach sexual maturity.
I agree the factory farming revolution has come to Africa—and its horrible, and it shouldn’t be acccepted for development. The church organisation I work with is currently teaching people to intensively farm pigs and chickens which I think is horrible, but haven’t spoken up against yet because I think I would be dismissed out of hand. In towns and even rural centers factory farmed chickens that are bought as meat are becoming the norm which is super sad, a. Even in Northern Uganda here, most roadside chicken is factory/intensive-barn farmed and that’s a new development − 10 years ago fried street chicken was not common and mostly local chickens. Pigs are often farmed in stalls (which are net-negative IMO but aren’t as bad as western pig stalls). I struggle to find Beef and Dairy farming in Uganda which I personally think isn’t probably net positive for the animals—but I have a more rosy view of truly free-range animal’s lives than many.
Yes there’s a huge amount of work to be done to at the very least slow down the factory farming revolution here. Its going to be tough though, when almost no-one I’ve met here cares an Iota about the wellbeing of a chicken. The few times I’ve even tried to raise it get met with bemused looks at best.
Thank you for your detailed reply! I admire your courage to raise this issue in front of your colleagues/the locals there—I am not sure I would find the courage to do so.
I have some hope that there might at least be ways to reduce the % of factory farming there will be in poor countries in the world in the future. Some EAs are working on it and I am trying to see what I can help there too.
Nice point, Nick! Even considering the conditions are as bad as in high income countries (pessimistic), I Fermi estimated accounting for the meat-eater problem only decreases the cost-effectiveness of GiveWell’s top charities by 8.72 %. On the other hand, I did not account for future increases in the consumption of animals throughout the lives of people who are saved (optimistic), which usually follow economic growth. For reference, I also Fermi estimated the badness of the experiences of all farmed animals alive is 4.64 times the goodness of the experiences of all humans alive, which suggests saving a random human life results in a nearterm increase in suffering.
Thanks Vasco yeah I’ve read your great posts. I straight up disagree on the relative importance of farmed welfare to human welfare, but that’s because Im not a hedonistic utilitarian and have far lower moral weights for animals than RP.
Even if the current population isn’t consuming much factory-farmed meat, if it’s children’s lives being saved, the amount they consume over the next half century or so may be substantial as the countries develop and adopt more industrialised food production. Also, saving lives today seems likely to increase population in future (I recall a GiveWell-commissioned study on this), so potentially leading to greater factory-farmed meat consumption.
I feel like this answer to the problem is easily forgotten by me, and probably a lot of similar-minded people who post here, because it’s not a clever, principled philosophical solution. But on reflection, it sounds quite reasonable!
I like this explanation a lot, from an 80000 Hours podcast:
This is also how I think about the meat eater problem. I have a lot of uncertainty about the moral weight of animals, and I see funding/working on both animal welfare and global development as a compromise position that is good across all worldviews. (Your certainty in the meat eater problem can reduce how much you want to fund global development on the margin, but not eliminate it altogether.)
Thanks for sharing that, Karthik.
Note stopping to support global health and developnent is not on the table. We are discussing marginal donations/spending. The global spending on global health and development would still be much larger than that on animal welfare after a small donor or even Open Phil changed to overwhelmingly supporting animal welfare.
Imagine you had to choose between donating 1 k$ to The Humane League (THL) or x k$ to GiveWell’s top charities. How large would x have to be for you to pick the latter? I would be curious to know your reasoning too. I guess you reject expected value maximisation. For reference, I estimated corporate campaigns for chicken welfare are 1.51 k times as cost-effective as GiveWell’s top charities.
The amount of global spending on each cause is basically irrelevant if you think most of it is non-impactful. Imaginine that John Q Warmglow donates $1 billion to global health, but he stipulates that that billion can only be spent on PlayPumps. Then global spending on GHD is up by $1 billion, but the actual marginal value of money to GHD is unchanged, because that $1 billion did not go to the best opportunities, the ones that would move down the marginal utility of money to the whole cause area. I understand you’re aware of this, which is why your Fermi estimates focus on the marginal value of money to each cause by comparing the best areas within each cause. But the level of global spending on a cause contributes very little to the marginal value of money if most of that spending is low-impact.
I don’t have a satisfying answer to what x is for me. I will say somewhere between 0.5 and 1.5, corresponding to the intuition that neither GHD nor FAW dominates each other. I would guess my cruxes with you come from two sources:
My median moral weight on chickens is much less than 0.33, ~2 OOMs less.[1] This is a difficult inferential gap to cross.
I think the quality of FAW cost-effectiveness estimates is vastly lower than GHD cost-effectiveness estimates, making the comparison apples-to-oranges. Saulius’s estimates are a good start on a hard problem, but
There are a lot of made-up numbers based on intuition (e.g. their assumption of 24% compliance with pledges in the absence of follow-up pressure is wildly out of line with my intuitions)
There’s likely steeply declining returns to effort given that campaigns will initially target the lowest hanging fruit, and eventually things will get much harder. Making a cost-effectiveness estimate based on early successful attempts is not representative of the value of future funding.
This is not a knock on people who are doing the best they can with limited data. I am just not comfortable taking these as unbiased estimates and I put a pretty high premium on having more certain evidence.
I see my views as consistent with expected utility maximization coupled with risk aversion, but not as expected value maximization (which, as its commonly defined, implies risk neutrality). The more uncertainty you have about a cause area, the more a risk-averse decisionmaker will want to hedge. (Edit: I also really like this argument for having a preference for certainty.)
I understand RP is estimating welfare ranges rather than moral weights, but I think you have to do some sneaky philosophical equivalences to use them as weights in a cost-effectiveness estimate. I’m open to being wrong about that.
(I realise this was posted a month ago but) this sounds to me like it overstates how bad global health aid is? I think all GiveWell top charities are existing organisations and programs that GiveWell only advocates increasing spending to, so surely effective aid existed before GiveWell did. Moreover, I have a not-particularly-concrete impression that e.g. vaccine distribution is only not an EA cause because it was already fully funded (at least in the easy cases) by non-EAs, so that our top charities are very much “top remaining” and not “best ever”.
I have the impression that even if EA and OpenPhil collectively tomorrow decided to move all of our global health funding to animals, there would still be a lot of effective global development aid—there would still be e.g. Gavi and the Bill and Melinda Gates Foundation (which sure, does ineffective things, but does effective things too) and many others. Wouldn’t that still meet the need you identified in your original answer for a compromise position?
Thanks for the follow up!
Just to clarify, I only care about the marginal cost-effectiveness. However, I feel like some intrinsically care about spending/neglectedness independently of how it relates to marginal cost-effectiveness.
Note this also applies to animal welfare.
Thanks for explaining your views! Your moral weight is 1 % (= 10^-2) of mine[1], and I multiplied Saulius’ mainline estimate of 41 chicken-years per $ by 0.2[2]. So, ignoring other disagreements, your marginal cost-effectiveness would have to be 1.32 % (= 0.2/(1.51*10^3*0.01)) the non-marginal cost-effectiveness linked to Saulius’ mainline estimate for corporate campaigns for chicken welfare to be as cost-effective as GiveWell’s top charities. Does this sound right? Open Phil did not share how they got to their adjustment factor of 1⁄5, and I do agree it would be great to have more rigorous estimates of the cost-effectiveness of animal welfare interventions, so I would say your intuition here is reasonable, although I guess you are downgrading Saulius’ estimate too much.
On the other hand, I find it difficult to understand how one can get to such a low moral weight. How many times as large would your moral weight become conditioning on (risk-neutral) expected total hedonistic utilitarianism?
Thanks for clarifying. Given i) 1 unit of welfare with certainty, and ii) 10 x units of welfare with 10 % chance (i.e. x units of welfare in expectation), what is the x which would make you value i) as much as ii) (for me, the answer would be 1)? Why not a higher/lower x? Are your answers to these questions compatible with your intuition that corporate campaigns for chicken welfare are 0.5 to 1.5 times as cost-effective as GiveWell’s top charities? If it is hard to answer these questions, is there a risk of your risk aversion not being supported by seemingly self-evident assumptions[3], and instead being a way of formalising/rationalising your pre-formed intuitions about cause prioritisation?
I strongly endorse expected total hedonistic utilitarianism (here is your sneaky philosophical equivalence :), and I am happy to rely on Rethink Priorities’ median welfare ranges.
Since Open Phil thinks “the marginal FAW [farmed animal welfare] funding opportunity is ~1/5th as cost-effective as the average from Saulius’ analysis [which is linked just above]”.
I think it makes all sense to be risk averse with respect to money, but risk neutral with respect to welfare, which is what is being discussed here.
I want to be clear that I see risk aversion as axiomatic. In my view, there is no “correct” level of risk aversion. Various attitudes to risk will involve biting various bullets (St Petersburg paradox on the one side, concluding that lives have diminishing value on the other side), but I view risk preferences as premises rather than conclusions that need to be justified.
I don’t actually think moral weights are premises. However, I think in practice our best guesses on moral weights are so uninformative that they don’t admit any better strategy than hedging, given my risk attitudes. (That’s the view expressed in the quote in my original comment.) This is not a bedrock belief. My views have shifted over time (in 2018 I would have scoffed at the idea of THL and AMF being even in the same welfare range), and will probably continue to shift.
Yes, I am formalizing my intuitions about cause prioritization. In particular, I am formalizing my main cruxes with animal welfare—risk aversion and moral weights. (These aren’t even cruxes with “we should fund AW”, they are cruxes only with “AW dominates GHD”. I do think we should reallocate funding from GHD to AW on the margin.)
Is my risk aversion just a guise for my preference that GHD should get lots of money? I comfortably admit that my choice to personally work on GHD is a function of my background and skillset. I was a person from a developing country, and a development economist, before I was an EA. But risk aversion is a universal preference descriptively – it shouldn’t be a high bar to believe that I’m actually just a risk averse person.
At the end of the day, I hold the normie belief that good things are good. Children not dying of malaria is good. Chickens not living in cages is good. Philosophical gotchas and fragile calculations can supplement that belief but not replace it.
Thanks for clarifying.
Are you saying that you are more likely than not to update towards animal welfare, or that you expect to update towards animal welfare? The former is fine. If the latter, it makes sense for you to update all the way now (one should not expect future beliefs to differ from past beliefs).
Nice to know.
One could work in a certain area, but support moving marginal donations from that area to animal welfare[1], as you just illustrated:
Thanks for being transparent about this! I think it would be good for more people like you, who do not think spending on animal welfare should increase a lot, to clarify what they believe is more cost-effective at the margin (as this is what matters in practice).
Right, but risk aversion with respect to resources makes sense because welfare increases sublinearly with resources. I assume people are less risk averse with respect to welfare. Even if people are significantly risk averse with respect to welfare, I do not think we should elevate this to being normative. People also discount the welfare of their future selves and foreigners. People in and governments of high income countries could argue they are already doing something pretty close to optimal with respect to supporting people in extreme poverty given their decriptive preferences. This may be right, but I would say such preferences are misguided, and that they should be much more impartial with respect to nationality.
I think the vast majority of people arguing that animal welfare should receive way more funding would agree with the above. I certainly do. I just do not think the calculations are fragile to the extent that the current porfolio can be considered anywhere close to optimal. I Fermi estimated buying organic eggs is 2.11 times as cost-effective as donating to GiveWell’s top charities[2], and I think that is far from the most cost-effective interventions in the space. which is what you suggest based on your guess that corporate campaigns for chicken welfare are 0.5 to 1.5 times as cost-effective as GiveWell’s top charities.
I support increasing donations to animal welfare, but I have not been paid for my work in the area. Personal fit plays much less of a role in deciding donations than in deciding jobs. It still plays some role because one could be better suited to assess donation opportunities in some areas.
My estimate relies on Rethink Priorities’ median welfare range for chicken, but it does not make any use of Saulius’ estimates, which are one of your 2 major sources of scepticism.
Hi Sammy, I’ve also found the meat eater problem compelling/distressing. One idea that gives me peace is an “offsetting” argument—that the most effective animal welfare charities use resources remarkably well. I’ve seen estimates online that the average American eats 7,000 animals over the course of their life (98.5% of which are fish and chicken). My intuition is that the average recipient of a GiveWell-recommended program eats slightly fewer animals, though I could be wrong.
The Humane League claims to be able to be able to spare a hen from the life of extreme confinement for just $2.63, and the Fish Welfare Initiative claims to be able to help a fish live in lower stocking density, higher water quality environments for just $1. Altogether, this suggests that donating ~$11,000 to cost-effective farmed animal welfare organizations mitigates a substantial amount of the harm that a meat-eater might cause (consuming 4500 fish and 2400 chickens). So a portfolio of global health and animal welfare donations seems likely to do good across both worldviews.
A few other scattered thoughts:
Saving human lives doesn’t just contribute to the problem of animal consumption, I hope it it also accelerates the solutions to factory farming.
The extent to which averting a death of a child under 5 (which GiveWell programs primarily target) increases total global population is unclear to me. Families that lose a child may be more inclined to have another baby than families that don’t. My guess is that lifesaving GiveWell-style charities do increase population in the short term, but not by one full person per life saved.
I would recommend this blog on the idea of a “moral parliament”, which I think can be an interesting thought exercise for resolving tensions like this. Rethink Priorities also has a cool tool for this idea.
I know this may not be fully satisfying, and this isn’t a strong argument against using your resources to go all in on animal welfare (which I think is a great thing to do), but I hope it might be helpful.
Yes, sufficient donation to animal welfare can make it net positive. But it doesn’t sound so good when one draws out the impact matrix:
Donate only to animal charities: +100
Donate only to human charities: −10
(not −100, because animal welfare is neglected, so proactive work on it does more good than creating more passive animal-eaters does bad)
Donate half to both: +45
(edited to add: If one thinks the animal eater problem’s conclusions are more likely true, such that these numbers could represent an average of one’s probability distribution. It looks like there’s disagreement in values about whether it’s valuable itself to for an action to be {at least positive} per se; I write about this extensively below)
That has the same structure as this:
Donate only to EA charities: +100
Donate only to bad-thing-x: −10
Not naming a particular bad-thing, because it’s unnecessary, but you can imagine.
Donate half to both: +45
If these feel different, a relevant factor may be how uplifting humans isn’t the kind of thing that narratively should be bad. It’s a central example of something many feel is supposed to be good—improving or saving lives. And in a different world, it would be good, because humans are not inherently evil.
I would point to purchase fuzzies and utilons separately in situations like this. If one believes the median human life has a net-negative impact, and also cares a lot about humans, and feels compelled to ‘act on’ their care for humans somehow, there are various less-harmful or net-positive ways they could. Here are some ways from brainstorming:
Find effective positive-net-impact things to do that involve ‘helping humans’ instrumentally
Example: Try to figure out the sources of mental suffering among alignment researchers, and potential ways to better this situation. (Or donate to this kind of work)
Same but for other EA cause areas
Help humans in small-scale but emotionally resonant ways
Helping a community or friends, separately from EA focus
Find some human charities that are maybe less effective at saving humans, but also a lot less likely to be net-negative
This comment has some examples
Help humanity become more morally reflective rather than trying to help lives directly
Work on alignment to help humans and other animals in the long-term future (!)
This is unresponsive to (what I perceive as) the best version of Sam’s argument, which is that a portfolio approach does more good given uncertainty about the moral weight on animals. Your impact matrix places all its weight on the view that animals a high enough moral value that donating to humans is net negative.
If you have a lot of uncertainty and you are risk averse, then a portfolio approach is the way to go. If you believe that there is a near 100% chance that helping poor people is bad for the world, then sure, don’t try the portfolio approach. But that’s a weirdly high amount of certainty, and I think you should question the process that led you there.
No, this is totally wrong. Whatever your distribution of credences of different possible moral weights of animals, either the global health charity or the animal welfare charity will do more good than the other, and splitting your donations will do less good than donating all to the single better charity.
This is why I said risk aversion matters—see this for a detailed explanation. Or see the back and forth with quila that inspired me to post it
Risk aversion doesn’t change the best outcome from donating to a single charity to splitting your donation, once you account for the fact that many other people are already donating to both charities.
Given that both orgs already have many other donors, the best action for you to take is to give all of your donations to just one of the options (unless you are a very large donor).
Yes, my response is from the perspective of the EA movement rather than any individual
If by weight you meant probability, then placing 100% of that in anything is not implied by a discrete matrix, which must use expected values (i.e the average of {probability × impact conditional on probability}). One could mentally replace each number with a range for which the original number is the average.
(It is the case that my comment premises a certain weighting, and humans should not update on implied premises, except in case of beliefs about what may be good to investigate, to avoid outside-view cascades.)
I think beliefs about risk-aversion are probably where the crux between us is.
Uncertainty alone does not imply one should act in proportion to their probabilities.[1]
I don’t know what is meant by ‘risk averse’ in this context. More precisely, I claim risk aversion must either (i) follow instrumentally from one’s values, or (ii) not be the most good option under one’s own values.[2]
Example of (i), where acting in a way that looks risk-averse is instrumental to fulfilling ones actual values: The Kelly criterion.
In a simple positive-EV bet, like at 1:2-odds on a fair coinflip, if one continually bets all of their resources, the probability they eventually lose everything approaches 1 as all their gains are concentrated into an unlikely series of events, resulting in many possible worlds where they have nothing and one where they have a huge amount of resources. The average resources had across all possible worlds is highest in this case.
Under my values, that set of outcomes is actually much worse than available alternatives (due to diminishing value of additional resources in a single possible world). To avoid that, we can apply something called the Kelly criterion, or in general bet with sums that are substantially smaller than the full amount of currently had resources.
This lets us choose the distribution of resources over possible worlds that our values want to result from resource-positive-EV bets; we can accept a lower average for a more even distribution.
Similarly, if presented with a series of positive-EV bets about things you find morally valuable in themselves, I claim that if you Kelly bet in that situation, it is actually because your values are more complex than {linearly valuing those things} alone.
As an example, I would prefer {a 90% chance of saving 500 good lives} to {a certainty of saving 400} in a world that already had many lives, but if those 500 lives were all the lives that exist, I would switch to preferring the latter—a certainty of only 100 of the 500 dying—even if the resulting quantities then became the eternal maximum (no creation of new minds possible, so we can’t say it actually results in a higher expected amount).
This is because I have other values that require just some amount of lives to be satisfied, including vaguely ‘the unfolding of a story’, and ‘the light of life/curiosity/intelligence continuing to make progress in understanding metaphysics until no more is possible’.
Another way to say this would be to say that our values are effectively concave over the the thing in question, and we’re distributing them across possible futures.
This is importantly not what we do when we make a choice in an already large world, and we’re not effecting all of it—then we’re choosing between, e.g., {90%: 1,000,500, 10%: 1,000,000} and {100%: 1,000,400}. (And notably, we are in a very large world, even beyond Earth.)
At least my own values are over worlds per se, rather than the local effects of my actions per se. Maybe the framing of the latter leads to mistaken Kelly-like tradeoffs[3], and acting as if one assigns value to the fact of being net-positive itself.
(I expanded on this section about Kelly in a footnote at first, then had it replace example (i) in the main post. I think it might make the underlying principle clear enough to make example (ii) unnecessary, so I’ve moved (ii) to a footnote instead.)[4]
There are two relevant posts from Yudkowsky’s sequences that come to mind here. I could only find one of them, ‘Circular Altruism’. The other was about a study wherein people bet on multiple outcomes at once in proportion to the probability of each outcome, rather than placing their full bet on the most probable outcome, in a simple scenario where the latter was incentivized.
(Not including edge-cases where an agent values being risk-averse)
It just struck me that some technical term should be used instead of ‘risk aversion’ here, because the latter in everyday language includes things like taking a moment to check if you forgot anything before leaving home.
Example of (ii), where I seem to act risk-unaverse
I’m offered the option to press a dubious button. This example ended up very long, because there is more implied uncertainty than just the innate chances of the button being of either kind, but maybe the extra detail will help show what I mean / be more surface for a cruxy disagreement to be exposed.
I think (66%) it’s a magic artifact my friends have been looking for, in which case it {saves 1 vegan[5] who would have died} when pressed. But I’m not sure; it might also be (33%) a cursed decoy, in which case it {causes 1 vegan[5] who would not have died to die} when pressed instead.
I can’t gain evidence about which possible button it is. I have only my memories and reasoning with which to make the choice of how many times to press it.
Simplifying assumptions to try to make it closer to a platonic ideal than a real-world case (can be skipped):
The people it might save or kill will all have equal counterfactual moral impact (including their own life) in the time which would be added or taken from their life
Each death has an equal impact on those around them
The button can’t kill the presser
These are unrealistic, but they mean I don’t have to reason about how at-risk vegans are less likely to be alignment researchers than non-at-risk vegans who I risk killing, or how I might be saving people who don’t want to live, or how those at risk of death would have more prepared families, or how my death could cut short a series of bad presses, anything like that.
In this case, I first wonder what it means to ‘save a life’, and reason it must mean preventing a death that would otherwise occur. I notice that if no one is going to die, then no additional lives can be saved. I notice that there is some true quantity of vegans who will die absent any action, and I would like to just press the button exactly that many times, but I don’t know that true quantity, so I have reason about it under uncertainty.
So, I try to reason about what that quantity is by estimating an amount of lives at various levels of at-risk; and though my estimates are very uncertain (I don’t know what portion of the population is vegan, nor how likely different ones are to die), I still try.
In the end I have a wide probability distribution that is not very concentrated at any particular point, and which is not the one an ideal reasoner would produce, and because I cannot do any better, I press the button exactly as many times as there are deaths in the distribution’s average[6].
More specifically, I stop once it has a ≤ 50% chance of {saving an additional life conditional on it already being a life-saving button}, because anything less, when multiplied by the 66% chance of it being a life-saving button, would be an under 33% total chance of saving a life compared to a 33% chance of certainly ending one. The last press will have only a very slightly positive EV, and one press further would have a very slightly negative EV.
Someone following a ‘risk averse principle’ might stop pressing once their distribution says an additional press scores less than 60% on that conditional, or something. They may reason, “Pressing it only so many times seems likely to do good across the vast majority of worldviews in the probability distribution,” and that would be true.
In my view, that’s just accepting the opposite trade: declining a 60% chance of preventing a death in return for a 40% chance of preventing a death.
I don’t see why this simple case would not generalize to reasoning about real-world actions under uncertainties about different things like how bad the experience would be as a factory farmed animal. But it would be positive for me to learn of such reasons if I’m missing something.
(To avoid, in the thought experiment, the very problem this post is about)
(given the setup’s simplifying assumptions. in reality, there might be a huge average number that mostly comes from tail-worlds, let alone probable environment hackers)
I agree that uncertainty alone doesn’t warrant separate treatment, and risk aversion is key.
(Before I get into the formal stuff, risk aversion to me just means placing a premium on hedging. I say this in advance because conversations about risk aversion vs risk neutrality tend to devolve into out-there comparisons like the St Petersburg paradox, and that’s never struck me as a particularly resonant way to think about it. I am risk averse for the same reason that most people are: it just feels important to hedge your bets.)
By risk aversion I mean a utility function that satisfies u(E[X])>E[u(X)]. Notably, that means that you can’t just take the expected value of lives saved across worlds when evaluating a decision – the distribution of how those lives are saved across worlds matters. I describe that more here.
For example, say my utility function over lives saved x is u(x)=√x. You offer me a choice between a charity that has a 10% chance to save 100 lives, and a charity that saves 5 lives with certainty. The utility of the former option to me is u(x)=0.1⋅√100=1, while the utility of the latter option is u(x)=1⋅√5. Thus, I choose the latter, even though it has lower expected lives saved (E[x]=0.1⋅100=10 for the former, E[x]=5 for the latter). What’s going on is that I am valuing certain impact over higher expected lives saved.
Apply this to the meat eater problem, where we have the choices
spend $10 on animal charities
spend $10 on development charities
spend $5 on each of them
If you’re risk neutral, 1) or 2) are the way to go – pick animals if your best bet is that animals are worth more (accounting for efficacy, room for funding, etc etc), and pick development if your best bet is that humans are worth more. But both options leave open the possibility that you are terribly wrong and you’ve wasted $10 or caused harm. Option 3) guarantees that you’ve created some positive value, regardless of whether animals or humans are worth more. If you’re risk-averse, that certain positive value is worth more than a higher expected value.
It sounds like we agree about what risk aversion is! The term I use that includes your example of valuing the square root of lives saved is a ‘concave utility function’. I have one of these, sort of; it goes up quickly for the first x lives (I’m not sure how large x is exactly), then becomes more linear.
But it’s unexpected to me for other EAs to value {amount of good lives saved by one’s own effect} rather than {amount of good lives per se}. I tried to indicate in my comment that I think this might be the crux, given the size of the world.
(In your example of valuing the square root of lives saved (or lives per se), if there’s 1,000 good lives already, then preventing 16 deaths has a utility of 4 under the former, and √1000−√984 under the latter; and preventing 64 is twice as valuable under the former, but ~4x as valuable under the latter)
Your parenthetical clarifies that you just find it weird because you could add a constant inside the concave function and change the relative value of outcomes. I just don’t see any reason to do that? Why does the size of the world net of your decision determine the optimal decision?
The parenthetical isn’t why it’s unexpected, but clarifying how it’s actually different.
As an attempt at building intuition for why it matters, consider if an agent applied the ‘square of lives saved by me’ function newly to each action instead of keeping track of how many lives they’ve saved over their existence. Then this agent would gain more utility by taking four separate actions, each of which certainly save 1 life (for 1 utility each), than from one lone action that certainly saves 15 lives (for 3.87 utility). Then generalize this example to the case where they do keep track, and progress just ‘resets’ for new clones of them. Or the real-world case where there’s multiple agents with similar values.
I describe this starting from 6 paragraphs up in my edited long comment. I’m not sure if you read it pre- or post-edit.
I suppose that is a coherent worldview but I don’t share any of the intuitions that lead you to it.
Could you describe your intuitions? ‘valuing {amount of good lives saved by one’s own effect} rather than {amount of good lives per se}’ is really unintuitive to me.
To me, risk aversion is just a way of hedging your bets about the upsides and downsides of your decision. It doesn’t make sense to me to apply risk aversion to objects that feature no risk (background facts about the world, like its size). It has nothing to do with whether we value the size of the world. It’s just that those background facts are certain, and von Neumann-Morgenstern utility functions like we are using are really designed to deal with uncertainty.
Another way to put it is that concave utility functions just mean something very different when applied to certain situations vs uncertain situations.
In the presence of certainty, saying you have a concave utility function means you genuinely place lower value on additional lives given the presence of many lives. That seems to be the position you are describing. I don’t resonate with that, because I think additional lives have constant value to me (if everything is certain).
But in the presence of uncertainty, saying that you have a concave utility function just means that you don’t like high-variance outcomes. That is the position I am taking. I don’t want to be screwed by tail outcomes. I want to hedge against them. If there were zero uncertainty, I would behave like my utility function was linear, but there is uncertainty, so I don’t.
This is so interesting to me.
I introduced this topic and wrote more about it in this shortform. I wanted to give the topic its own thread and see if others might have responses.
I do this too, but even despite the worlds size making my choices mostly only effecting value on the linear parts of my value function! Because tail outcomes are often large. (Maybe I mean something like kelly-betting/risk-aversion is often useful to fulfill instrumental subgoals too).
(Edit: and I think ‘correctly accounting for tail outcomes’ is just the correct way to deal with them).
Yes, though it’s not because additional lives are less intrinsically valuable, but because I have other values which are non-quantitative (narrative) and almost maxxed out way before there are very large numbers of lives.
A different way to say it would be that I value multiple things, but many of them don’t scale indefinitely with lives, so the overall function goes up faster at the start of the lives graph.
It’s a tough question and something I’ve tried to wrap my head around as well. All of the threads in the comments here are quite helpful!
This point you’ve made, Sam, is also something I have thought about:
Awareness of animal welfare issues tends to increase as people get richer and have more space to think about something other than their immediate needs. Of course, factory farming is worse in richer societies, but I think those societies are also closest to overcoming factory farming / the worst farming practices (veganism is more popular, bans of cages, mandatory pre-slaughter stunning, R&D into alt proteins, etc.). This hinges on a few assumptions, which can be debated, but I tend to find it plausible.
That said, I still exclusively support animal charities at this stage, since I think they are anyway far more cost-effective at improving sentient lives (see the points made by Vasco Grilo and Ben Millwood made in this thread).
Brian Tomasik has argued that if (a) wild animals have negative welfare on net, and (b) humans reduce wild animal populations, then that may swamp even the horrific scale of factory farming.
I personally think the meat eater problem is very serious, and the best way around it is to just donate to effective animal welfare charities! Those donations would be orders of magnitude more cost-effective than the best human-centered alternatives.
Nice question, Sammy! I worry the meat-eater problem is mostly a distraction. If one values 1 unit of welfare in animals as much as 1 unit of welfare in humans, and does not think Rethink Priorities’ welfare ranges are wildly off, the best animal welfare interventions will be much more cost-effective than the best interventions to save human lives. I estimated corporate campaigns for chicken welfare, such as the ones supported by The Humane League (THL), are 1.51 k times as cost-effective as GiveWell’s top charities.
Its not only the welfare ranges, but also assuming we are hedonistic utilitarians which is pretty important and I would be interested to know how man EAs are in that boat. In general though your point stands!
Thanks, Nick.
Rejecting this implies rejecting hedonistic utilitarianism if welfare above is interpreted as hedonic welfare. I have now changed “welfare” to “hedonic welfare” above.
Holding everything else constant in my calculations, one would have to be less than 0.0662 % (= 1/(1.51*10^3)) hedonistic utilitarian to prioritise GiveWell’s top charities over corporate campaigns for chicken welfare. Weighting hedonistic utilitarianism so lightly, I think:
Current factory-farming could easily be justified. In essence, because one would be able to value eating factory-farmed animals up to 1.51 k times as much as today.
There would also be no obvious reason to support global health and development interventions. Such support makes sense assuming an additional year of healthy life or a given relative increase in income are worth roughly the same regardless of the person, but this is only obviously the case if one values hedonic welfare the same regardless of who experiences it.
I suspect people who prioritise global health and development over animal welfare implicitly endorse hedonistic utilitarianism for comparisons between interventions targeting people, but reject it for comparisons across species.
Hi Vasco, I first learned about the meat eater problem from your post. Thank you for your insight.
My thoughts:
It’s not clear that saving someone’s life causes a net increase in the population.
There was some Givewell-commissioned research that did find that saving lives likely leads to future population increases. I imagine there’s a fair amount of uncertainty, but it seemed to be the best information available at the time I was looking into this a few years ago. I could dig it up if it’s of interest and difficult to find.
Yes there may be Givewell research saying that, but its still very unclear, and the mainstream public health view (for what its worth) has generally been that better healthcare and saving lives may well lead to lower fertility rates and lower populations in the medium/long term.
Are there any good research articles that do a decent job of isolating the role of reducing mortality rates? Review articles would be particularly useful.
Here’s a link to the GiveWell-commissioned research that I have: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3635855 .
I think if you put some weight on viewpoint pluralism you should mostly not conclude that other peoples’ lives aren’t valuable because those people will make the wrong moral choices.
Im getting at that most people would would not go out of their way to save baby Hitler. I value Hitler’s life, but I also wouldn’t save his life.
I don’t really understand this stance, could you explain what you mean here?
Like Sammy points out with the Hitler example, it seems kind of obviously counterproductive/negative to “save a human who was then going to go torture and kill a lot of other humans”.
Would you disagree with that? Or is the pluralism you are suggesting here specifically between viewpoints that suggest animal suffering matters and viewpoints that don’t think it matters?
As I understand worldview diversification stances, the idea is something like: if you are uncertain about whether animal welfare matters, then you can take a portfolio approach where with some fraction of resources, you try to increase human welfare at the cost of animals and with a different fraction of resources you try to increase animal welfare. The hope being that this nets out to positive in “world’s where non-human animals matter” and “world’s where only humans matter”.
Are you suggesting something like that or is there a deeper rule against “not concluding that the effects of other people’s lives are net negative” when considering the second order effects of whether to save them that you are pointing to?
I’m not proposing any sort of hard rule against concluding that some people’s lives are net negative/harmful. As a heuristic, you shouldn’t think it’s bad to save the lives of ordinary people who seem to be mostly reasonable, but who contribute to harmful animal agriculture.
The pluralism here is between human viewpoints in general. Very naively, if you think every human has equal insight into morality you should maximize the lifespan and resources that go to any and all humans without considering at all what they will do. That’s too much pluralism, of course, but I think refraining from cheaply saving human lives because they’ll eat meat is too far in the other direction.
This is where we need a broad perspective.
Long-term, we solve the problem of meat-eating with artificial protein, which also solves many other problems.
Medium-term, we work to end factory-farming, which needlessly increases the suffering of animals. (I don’t want to get into it because there are many experts here and I’m not one of them, but it may be arguable that an animal which is bred for food but gets to live a decent life in a field is better off than if it hadn’t been born because people didn’t need to eat it. However, in the case of factory-farming, such an argument seems totally untenable).
Short-term, we accept that we live in an imperfect world and that most people value saving human lives, even at the cost of animal lives. So we work to save human lives and improve health and improve quality of life, and instead of losing sleep over the calculation of the net impact on animals, we support the amazing organisations who are working to end factory-farming (like Farmkind) and to develop alternative protein (like GFI).
It’s valuable to discuss questions like this, and I absolutely do not claim to have a definitive answer—all I say is that when I think about this, that’s how I rationalise it.
This problem[1] was one I independently noticed before finding EA.
I think in some ways the actions implied under it are similar: focusing on long-term world-change so that the world is eventually organized by morality.
If ‘how do you deal with it’ means ‘how do you make your actions compatible with this view’, my answer is that I’m trying to help with {figuring out how we can make powerful AIs also be benevolent}.[2]
If ‘how do you deal with it’ means ‘how do you convince yourself it is false, or that things some EA orgs are contributing to are still okay given it’, I don’t think this is a useful attitude to have towards troubling truths.
I agree with this:
Many current human lives causing extreme suffering by funding factory farming, enough to outweigh the positive value of their own life
A nice fiction about coming into existence in a world with normalized moral catastrophe, and acting heroically: the sword of good by Yudkowsky
Well said and important :)
Good question. Does this also work in the opposite direction? ~ Worry less about catastrophic or existential risks because there’d be fewer animals in factory farms?
I have an intuition that any ASI that wipes out humans does the same to non-human animals though.
For a standard utilitarian, a benevolent superintelligence would create enough happiness (and not allow factory farming) to outweigh any current suffering due to the large length and size of the future.
For a suffering-focused altruist (such as myself), it’s not that simple, although in any case it mostly revolves around (i) the possibility of locally-originating long-term s-risks (rather than factory farming, if it ends near-term), and (ii) the ability of aligned ASIs to reduce s-events in unreachable parts of the world through acausal trade; see my shortform expected value of alignment over extinction for negative utilitarians
We need to keep in mind that we are effective altruists (or effective anything) to the extent that we are also respected members of our communities, not only of the EA community. We live in a world in which meat-eaters are much of the population. In this kind of world, the belief that the life of a meat-eater life has a net negative value would turn us into a cult of phanatic monsters. Or at least we would be perceived in this way. We would have no chance whatsoever of contributing to the end of cruelty. We need to stay humans, connected to all kinds of humans. The quantitative approach to morality is powerful, one of the best ideas I can think of. But it’s not the only tool we should use.
I deal with this problem by donating (almost) exclusively to animal welfare charities. Besides, I think that the suffering of animals reared for human consumption is much more intense than those of humans, in general.
More development may at least indirectly contribute to hastening ubiquitous lab-grown meat becoming economically cheaper than non-lab-grown meat.
A lot of uncertainty here because I have no idea how much (if at all) more development may cause this, but, if it does, it leads to fewer total moral atrocities.
I would probably believe the poorest most rural parts of Africa would not be able to contribute to lab-grown meat development before it is brought efficiently to market. Furthermore, these parts of the world would likely be the last to adopt lab-gorwn meat.
RCT-informed interventions focused on the poorest will not increase demand for factory farmed meat—only broad based economic growth will do this. So one solution is to focus on micro interventions targeted at the extreme poor.
Another solution is to support the alternative proteins sector in LMICs, which could enable some degree of “leapfrogging” factory farmed meat and reduce carbon emissions.
I read some of the answer and they were really interesting. I like to simplify so i will bring my modest simple answer and my advice :
I have a amateur answer but i am not an expert although this topic interest me a lot being interested in sustainability, agronomy and ea (here global health and animal farming) :
Firstly pro donating to animal farming :
if you believe in animal having a moral weight similar (1 − 10x) to humans it is probable that saving a human life might be negative compared to saving 100 − 1000 animals.
We know that our way of living is not sustainable. We also know that one reason why our lifestyle is not sustainable is animal farming.
Secondly pro donating to save a life :
Saving a life is good (its one of the first moral reflex if it happens in front of you and)
You never know what people will do with their life. They probably eat local non industrial products and dont cause much pollution if they are poor.
My opinion : Saving a human life will probably do a little bad because it causes the death of some animals (not necessarily factory farm) and causes pollutions—we are not sustainable (dependent of polluting fertilizer N,P,K). Saving animals will probably do a little good because we would be more sustainable. So I would focus on saving non-human animals because it will probably help all animals and the opposite is probably false.
Now my advice :
If you don’t know the answer find some informations. When you have your answer keep being open-minded. I think this topic could be researched a hole lifetime without covering every aspect of it. Keep in mind since morals are subjectives there is no right answer. But since this quesion is important you should research the subject. I believe collective intelligence is the best “moral objectivity” we could acheive if everybody is well informed.
Life is never black or white and full of other dilemnas. Saving a human life or other animals are not your only two options wich brings me to my last opinion : before doing one or the other we should focus on discussing this subject with the most people possible. Or at least make sure everybody knows animals suffer (especially factory farmed) and everybody knows we can save lives with 5000$
Two things to consider:
Animal suffering is currently invertly U-shaped in income.
The rise of lab-grown meat could surely dramatically increase the negative relationship between animal suffering and income at high-levels.
If you find this question truly compelling—if it’s more than just an intellectual challenge for you—I would suggest reconsidering vegetarianism. When expanding your empathy starts to hinder your basic empathy for humans, or when being a vegetarian makes you think of your aunt as a “meat-eater” rather than as a warm and kind person...
I acknowledge that I haven’t provided a strong justification for my answer, and I don’t know your full set of beliefs and experiences, so this is definitly not a judgment. However, I do strongly believe that the “aunt” argument is a valid one.
I don’t think that expanding compassion to animals leads to reduced compassion for aunts and other humans. It could make one’s choices benefit humaunts less, but that’s arguably desirable.
I do agree that when we find ourselves dispassionate or hateful towards others that’s a sign that we may have stepped wrong at some point in our journey to do the most good
This makes me think that countries who as of yet don’t have an entrenched factory farming lobby/industry would benefit advocacy groups similar to Shrimp Welfare Project (work in the reverent countries with stakeholders to improve the wellbeing of farmed animals).
I began wondering if any org was approaching this similar to SWP. There seem to be two EA groups working on this:
Animal Advocacy Africa
Education for Africa Animal Welfare