A new article went viral on Twitter today: Nan Ransohoff's "The Third Wave of American Philanthropy" (link). Worth reading first.
Nan is right about the shape of what's coming: hundreds of billions in new philanthropic capital, no ecosystem yet to absorb it, and a shortage of builders and organizations. I very much agree with that sentiment and the direction. More money, more people willing to start things, more urgency.
But the conclusion I draw is a little different. The new philanthropic wave shouldn’t go hunting for problems in completely new places: it should look harder within one it already named and set aside, within an all-encompassing field that history reduced to a single term too small to carry its gravity.
Far more of the new philanthropic wave should go to helping animals. It should go to the trillions of lives in all their variety and multitude. “Animal welfare” isn’t one solved issue to cross off: it’s where most of the sentience and suffering is. It covers the vast majority of moral patients, all of them suffering gravely in every corner of the world. The future is still incredibly grim; AI could impose even more suffering if we don’t get this right.
Animals are not one issue
EAs recognized the importance of animal suffering – factory farming, wild animal suffering – long before the rest of the world, which still has not really recognized its importance. That was the insight. And then, somehow, we just decided to group it all as one single issue. One moral cause area.
Chickens, fish, pigs, shrimp, insects… farmed and wild. Nonhuman animals are the most populous category of moral patients by orders of magnitude. Yet we have collapsed all of it into a single line item that competes for attention with everything else as if it were just one intervention among many.
The problems are staring right at us
This could be the path we go down in the name of flourishing: we simply lower the bar on cost-effectiveness. We go looking everywhere for things to do, finding problems to solve, manufacturing causes to fund. But are those really the best use of a once-in-a-generation wave of capital and talent?
If we still take scale and cost-effectiveness seriously, we have so, so many problems staring right at us. They just happen not to be happening to humans.
What we need is a lot more people working on:
- getting billions of chickens out of cages;
- getting tens of millions of mother pigs out of gestation crates;
- giving billions of ducks access to water;
- stopping hundreds of millions of frogs from being dismembered alive;
- preventing tens of billions of chickens from fracturing under their own weight;
- sparing hundreds of millions of rabbits whose feet are sliced raw by the wire cages;
- stopping trillions of feeder fish from being eaten alive;
- preventing hundreds of millions of mice and rats from a slow death suffocating on their own blood;
- preventing trillions of insects from being farmed in horrendous conditions (even here, I'm reducing the most numerous animal species on Earth into a single word).
Just some examples, and the list goes on and on and on. I'm listing these only to make them slightly more granular.
There are a lot more neglected animal populations that no one is doing anything about, and we should be looking for those problems (I left my job recently to do exactly this). A lot of new effort should go toward finding new ways to help all those neglected beings.
AI x Animals
Contrary to popular belief, many "animal people" in EA recognize just as much that the most critical work today lies in making sure AI goes well. But while it is easy to appreciate AI's grave implications, what is even more alarming is something the discourse rarely touches: AI's implications for animals might still be graver.
We could flourish while the moral atrocity that is factory farming is left intact, or even expanded by AI. Trillions of sentient lives could continue suffering completely unheeded as they stay almost entirely outside the moral circle.
Yet anything in that AI × animals intersection gets filed, yet again, under "animals."
"AI x animals" is a sub-bucket of a sub-bucket, rather than treated as central to the long-term future we keep saying we care about. If we are serious about a future with less suffering, then what AI entrenches or prevents for animals is not a footnote to that future. It may be most of it.
Stop looking past them
Under any reasonable moral weight model, animal consciousness and suffering far trump that of humans - today and very possibly long into the future.
Beyond making sure we don’t all die, stopping animal suffering should be our first priority. We should address this before moving on to civics, art, and flourishing.
The problems are staring right at us, and we continue to look past them. And now, with all the new energy pointed at finding new problems to solve, we are about to look past them harder, with more money and more talent than ever. Let’s not.
Megaprojects for animals (or an updated version perhaps, this list being from 2022) seems more pertinent than ever.
matthes' recommendation to "take ownership of the entire evidence pipeline" seems more plausible than ever in this context:
@Fai had some excellent and prescient ideas. Anything new come to mind, Fai?
Thanks for asking, here are some ideas I personally wish funders would consider at least investigating. The epistemic status of some of these ideas is not great, and I never attempted any robust analysis on the expected values of these potential interventions/causes, but I hope they are worth investigating.
Nudge to spend another hour on this and turn it into a post?
Thanks for the nudge! I have something really important coming up in July, possibly the most important thing I might do so far in my career. I will consider after that. Feel free to nudge me again in August!
Wow very cool, best of luck!
Thanks Fai. The most farmed fish on earth is likely feeder fish for mandarin fish, numbered in the trillions (~1B mandarin fish x fed 2-4k feeder fish each). We’re working on it at Myrias.
Yes feeder fish for mandarin fish is a big category. But my understanding is that many species' fry are used, including pond loaches (rarely though, I believe). I am not sure the majority of them are one species (i.e. mud carp).
Also, since we need to count fry to come up with a few trillion figure for feeder fish for mandarin fish, we also need to count pond loach numbers by the fry stage. Estimates of pond loach survival rates from fry to sellable fish vary widely, from 2% to 10%. Given that the number of pond loach slaughtered each year is roughly 10B, that's ~100B-500B pond loaches slaughtered each year.
Yeah, maybe mud carp or some other species is no.1, but I am guessing it is also possible pond loach is still no.1 or close.
Yeah, from what I heard it’s mostly mud carps (土鲮), which are likely in the trillions. FWIW, pond loaches contain many different species too. Anyhow, both very numerous and neglected, deserving of some serious effort.
Yes, thanks for the reminder. I have long (incorrectly) thought pond loach is just one species, until Ryan pointed out that there are at least 4 (but seems like only two are commercially popular).
From what I learned, even though mud carp should be the biggest used fry for mandarin fish feed, many other species such as other carps and tilapia are also used in significant amounts. But in terms of cause priortization/conceptualization, grouping them together makes perfect sense!
Thanks, Fai. I like your ideas. Virtual Control Groups and LLM-agents are especially interesting to me right now. I want to look into the state of digital twins for various animals. This could not only obviate some animal testing, but also facilitate better translational medicine between humans, farmed animals, companion animals, and wild animals. It might also help us model levels of suffering and welfare improvement associated with interventions like novel pesticides, without the need for physical experimentation. Models would probably mostly cover individual physiology, but could also model population dynamics on farms or in the wild. Aware of anything on these fronts?
Also, your comment on pond loaches reminded me of our ~2021 discussions around animals in the long term future in space. I am planning on revisiting some of those topics.
I strongly agree that the animal welfare implications of AI should be owned at least as much by the AI safety space as by the animal welfare space, not that there needs to be a hard distinction between the two but there is obviously some declarative truth to it. Animal suffering is among the greatest lock-in risks.
I'm worried that many people outside the AW space believe the end of factory farming is a foregone conclusion. At one forum in SF in February mixing leaders from AI safety and AW, many AIS folks came away at least partly convinced by the AW folks that this cannot be taken for granted. AIS needs to take seriously the inside view of AW leaders that AI will not necessarily solve FAW by default, not to mention WAW.
Please do not write clickbait titles like this. Put the subject matter in the title.
Thanks for the suggestion
New title is better!
Sigh
I agree that animal welfare is underfunded, but who is this post trying to convince? If funders should fund the most cost-effective interventions or cause areas, then the argument needs to be made on the merits, which are not covered here. Animals don't just automatically win because there is lots of animal suffering today (unless maybe you are a fairly near-termist negative utilitarian).
If you’d like to read some arguments, I argue here that the most cost-effective neartermist interventions are in animal welfare. If you lean longtermist, I argue here that under many EAs’ risk aversion, marginal animal welfare donations still make more sense than marginal AI safety funding. If you’re a pure total utilitarian, I would still argue that direct efforts to improve the future for all sentient beings (future-oriented digital minds/animal welfare work) are plausibly higher EV even than x-risk reduction.
(I have already read these posts and much prefer them over the post we're commenting on here.)
I also recommend engaging with RP's new cross-cause work here. Curious to hear takes.
Well said. On top of this, many new interventions are going to be primarily talent constrained, whereas Animal Welfare is primarily funding constrained. Animal Welfare should be prioritized at least until funding is no longer the main constraint.
Thank you for the post!
I have long wished to write or see an EA forum post about this. Thanks for doing it!
I really like this framing, especially the push against “problem hunting” when there’s already an overwhelming amount of neglected suffering in front of us. Animals are not “one issue” but instead, the majority of moral patients. This is underappreciated, even within EA.
What struck me about the original article was the idea that we don’t yet have the ecosystem to absorb a huge influx of capital. I think that applies here too. Even if we take the case for prioritising animals as seriously as we can, it’s not obvious the field is currently set up to productively deploy funding at the scale you’re pointing to (multiple interventions, geographies, and organisations) without hitting bottlenecks or diminishing returns.
That seems less like a reason not to prioritise animals, and more like a reason to invest heavily in building out the space itself. The animal space doesn't have nearly enough people, enough organisations, and needs better-developed intervention areas, and especially more work on neglected animal populations (which is why what we're doing now at Myrias feels particularly valuable).
Overall, I agree with the core concern. There’s a real risk that a wave of new funding ends up looking for new, legible problems while continuing to underinvest in ones that are already vast, tractable, and unsolved. Making sure that doesn’t happen for animals seems like one of the key challenges for this next phase.
I don't understand the negative karma for a comment that has 5 agreement votes to 0 disagreement. I wonder if some of you can explain the reason for agreeing but downvoting this?
I downvoted it because it is AI generated, and reads (to me) like someone has prompted an AI to 'respond to this post', i.e. it doesn't really make any specific points. I could be wrong though, and clearly others have found value in it.
Hi Toby, that's a fair comment. I dictated a reply to the post to my LLM and then it put together into something less rambley and slightly more cohesive. I then edited it manually, but I guess some of the AI-stink remained! I do think I added something new to it, as my professional experience as an animal rights activist is that the animal space is quite under developed. So content was all mine (a human with experience in the field, albeit mostly agreeing with both pieces!), but the formatting and copywriting was handled by AI.
Makes sense Carlos! It might be a personal thing but when I see AI writing I often struggle to read it/ understand what is the idea of the author and what is the AI-fluff.
Great write up! I especially appreciate your one-line descriptions of some of the largest AW issues. Are you aware of any kind of database where these issues are catalogued, and if so, could you point me to it? We have a long list of interventions, but having a long list of discrete welfare issues would be at least equally useful.
Very well said, it's so sad what happens to so many animals and like you said the list goes on and on. And the funny thing is, we are animals. If AI goes well for animals, that implies it goes well for us.
But we really don't want speciesist AI. If we have AI that has a moral circle based on species membership, or based on certain capabilities like intelligence, what happens when it gets to the point where we are no longer intelligent enough to qualify, or the only species that's morally relevant is the AI species?
FWIW I agree with the sentiment!
However, I think this post as it stands leans a lot on the Importance dimension of the ITN framework, gestures at Neglectedness, and doesn't directly address Tractability. To be clear, that's fine! Not every post needs to address everything; some of the linked posts address those somewhat; other resources out there do too.
But I do think that the question of 'what can be done with 100s of extra $bn' is also a question of T and N, and I think part (majority?) of what Ransohoff is saying is that we're woefully unprepared ~1 level down from 'cause area' at the infrastructure / intervention / ecosystem / etc. level (where T and N mostly cash out, IMO).
If people have or can point to write-ups about how many and what size of shovel-ready projects we have for the various animal problems listed: definitely post and signal boost those!!
Maybe a crux here might be 'what sort of net effect one expects from more or less just spending any amount of $ to hire anyone you can to try to find those projects'. Not saying that's the author's take per se; but I am somewhat skeptical of things shaped like "we need[...] a lot more people working on [X]" if what they mean is 'we need to get warm bodies on this at all costs'. People can have net negative impact! There are opportunity costs to going all-in like that!
Overall I hope a large chunk of this wave of philanthropy goes to animal welfare, but also that it can be spent well and deployed quickly :)
E.g. mo's megaprojects link in another comment!
This lands for me, and the line I keep returning to is the one where you catch yourself reducing the most numerous animals on Earth to a single word. That instinct to collapse is the whole problem in miniature.
Think about how we treat human health by comparison. We take one species and split it into twenty urgent priorities: heart disease, cancers, maternal mortality, road accidents,criminal justice , governance... We refuse to flatten ourselves, because every facet of our suffering feels vivid and worth its own field. With animals we run the operation in reverse. We take thousands of species, or at least the dozens we treat most negligently, and fold them into one comparable unit, then ask that single unit to compete for attention against everything else. The asymmetry in how we draw the lines tracks nothing about the underlying scale of suffering. It tracks only how much we let ourselves care.
And the reason is a predictable bias, not a mystery. We value the pain of beings who look and sound like us, and our concern falls off sharply with distance: from our race, our gender, our nation, our species. It's scope insensitivity pointed at the moral circle itself. The chickens fracturing under their own weight and the feeder fish eaten alive are not suffering less than we would. People have tried to put numbers on this, and even the conservative welfare estimates imply aggregate suffering that dwarfs almost anything else we fund. The beings are just far enough from us that the figures don't move us the way they should. Their ordinary day likely holds more pain than our worst one.
This is where your AI point matters more than the discourse admits. The moral circle we draw right now is the training example a more capable intelligence learns from. If we build systems while modeling the lesson that distance and unfamiliarity justify discounting a being's pain, we are encoding the template that gets applied to us the moment we are the less powerful ones. How we treat what we can ignore today is the precedent for how we get treated when we can be ignored. That is not a footnote to the long-term future. As you say, it may be most of it.
Thank you for refusing to look past them.
I largely agree, the plight of animals is being looked past ... as is the only somewhat realistic solution to factory farming (and, therefore, the implications of AI on factory farming).
All or most funding currently allocated towards animal welfare in EA ought to be focused on but one intervention, in my view: policy advocacy to help get cultivated meat on supermarket shelves ASAP. No other intervention offers comparable upside (in terms of reduced suffering) of even a modest market penetration of cultivated meat.
I don't see another way to end factory farming in our lifetimes.
Just one person's strongly held view, though I'm also pleased to see Bruce Friedrich's talk at EAG next week titled 'the only tractable solution to industrial animal farming'. See some of you there!
Once the Bacon Bill is cooked, there may be an interesting piece for an unusual coalition to write on applying the proposed standards for pig lives for how AI safety may play out for humans - no minimum standards, commercial incentives, etc.
Wow, thanks for writing this. This article reminded me of the idea that an LLM token can carry vastly different levels of importance depending on the context. I think of A VLM's reasoning trace across analysis of factory farm cams (which could yield "ready_for_slaughter" entries into an Excel spreadsheet). Something that scares me about token cost reduction. I don't hear this talked about much at all.