Summary
The Alliance to Feed the Earth in Disasters (ALLFED) is an EA-aligned charity with potential for high cost effectiveness in the global poverty and existential risk spaces.
We have recently been awarded $70,000 by one of the EA Lotteries; however, considerable room exists for more funding. Should you consider donating to ALLFED, you can double your donation through the Facebook/Paypal Giving Tuesday matching event on November 27th (please see here for instructions and here to donate).
This post is a 2018 progress report building on the EA forum post introducing ALLFED in 2017.
I have also posted on the EA forum before about getting prepared for alternate foods (roughly those not dependent on sunlight that exploit biomass or fossil fuels) for agricultural catastrophes such as nuclear winter. This could save expected lives in the present generation for $0.20 to $400. These catastrophes have a number of routes to far future impact including loss of civilization and non-recovery, making other catastrophes more likely (e.g. totalitarianism), or worse values ending up in artificial general intelligence (AGI). In a recent EA forum post, I made the case that spending $100 million on alternate foods would likely be better cost-effectiveness than AI safety from a far future perspective.
ALLFED has an experienced team and board. With a small budget, it has achieved a significant amount this year, including five peer reviewed papers, a catastrophe planning session, and a dozen presentations. It has plans to increase preparedness with targeted planning and research. It has several volunteers who could contribute more if paid and is in general funding constrained. I have donated half my income the last three years to ALLFED. ALLFED has tax-free status in the US, and in the UK through CEA. I outline what could be achieved with different levels of funding and other ways to help.
Disclaimer/Acknowledgements
We would like to acknowledge our donors including CEA for funding the EA grant. Also thanks to Avi Norowitz and William Kiely for coordinating the EA Giving Tuesday effort. Sonia Cassidy and Finan Adamson contributed to the post. Opinions are my own and this is not the official position of ALLFED, CEA, Open Philanthropy Project, nor the Global Catastrophic Risk Institute.
Background and cost effectiveness of cause area
For more details, see here. The quick summary is that there is significant risk of nuclear war and other catastrophes that could dim or block the sun and dramatically reduce agricultural output. Previous work on storage or preventing war has hit diminishing returns. Alternate foods is a neglected solution that shows high cost effectiveness both from the perspective of the present generation and the long-term future. The Alliance to Feed the Earth in Disasters (ALLFED) appears to be the only organization in this space. This post reviews the accomplishments of ALLFED in the last 12 months and outlines what ALLFED would do with additional funding.
ALLFED mission/vision
Mission: Increase the preparedness, readiness (knowledge, resources, technology) of world bodies, governments, corporations, NGOs/people to be able to feed everyone in the event of a global catastrophe.
Vision: Form an alliance of key people/willing participants working to develop capability to enable response to global disruption of food supply.
ALLFED team & board
Dr. Gorm Shackelford has joined us as a new board member, bringing with him all-important agricultural expertise. Dr. Shackleford is a research affiliate at the Cambridge Centre for the Study of Existential Risk (CSER) and a post-doctoral research associate in Zoology at Cambridge.
Finan Adamson (recent coordinator of EA Seattle) has also joined ALLFED as our first full time employee, having previously volunteered with us for a year under a grant from CEA.
Please see our website for the full current team and board.
Accomplishments in the last 12 months
Research: We have given a dozen presentations at several conferences/universities including EAG San Francisco (poster and talk), Information Science and Technology (ISAT) / Defense Advanced Research Projects Agency (DARPA): Study About Technology In Agriculturally Troubled Environments (SATIATE), EA Seattle, EA Oxford, FHI, CSER, Society for Risk Analysis conference, International Food Policy Research Institute, University College London Institute for Risk & Disaster Reduction, and twice at the University of Alaska Fairbanks.
We also got five peer reviewed journal articles accepted:
Food without sun: Price and life-saving potential goes over the current prices of various alternate foods and how many more lives could be saved by alternate foods than stored food.
Micronutrient Availability in Alternative Foods During Agricultural Catastrophes dives into the question of getting micronutrients from alternate foods. Most of our past work has focused on getting enough calories. In a disaster eating a wide range of foods or supplementing with essential vitamins is important to reduce disease and stay healthy.
A National Pragmatic Safety Limit for Nuclear Weapon Quantities argues that a country should have no more than 100 nuclear weapons. This is because if a country were to use them, even if there was no retaliation, the nuclear “autumn” that results would cause unacceptable damage to the country that launches the weapons.
Classification of Global Catastrophic Risks Connected with Artificial Intelligence argues that artificial intelligence can cause global catastrophes in several dozen different ways. One of them is by disrupting electricity, and therefore industrial civilization. This is one of the catastrophes on which ALLFED works.
Global Catastrophic and Existential Risks Communication Scale proposes a color-coded scale for the priority of different global catastrophic risks based on impact and probability. Artificial intelligence alignment was considered red. Nuclear war and approximately 10% global agricultural shortfalls, on which ALLFED focuses, were rated orange. The scale did not cover neglectedness, but if it did, then agricultural catastrophes would be on par with AI.
We co-authored with Alexey Turchin “Classification of Global Solutions for the AI Safety Problem,” which won one of four global top prizes in GoodAI's General Artificial Intelligence Challenge.
Effective Thesis: We put several dozen ALLFED-related effective theses on the effective thesis website.
Planning: We ran another catastrophe planning exercise, this time at EAG London. We also attended several other conferences, including Global Challenges Foundation in Stockholm, Sweden, and Volcanoes on Cities in Naples, Italy. We also spent 3 months in India and Sri Lanka developing research collaborations and subcontinental preparedness; we met Prof. Swaminathan (the father of the Green Revolution in India) and visited the International Crops Research Institute for the Semi-Arid Tropics.
Mistakes
We brought on a fundraiser to focus on mainstream fundraising outside of EA. This arrangement did not work out. However, it only lasted one month, limiting the loss.
Support this year
I gave half of my income again for this year, but I was joined by a number of other donors. These included Jacob Trefethen, Ben West, Greg Colbourn, and Adam Gleave. We applied for funding from half a dozen institutions, and some are still pending.
What different levels of additional funding could do in 2019
Our basic plans for 2019 are to continue research under the Centre for Effective Altruism grant, perform high-value experiments, and do high-value planning.
The immediate task for the EA grant is the cost-effectiveness from a far future perspective of ameliorating a different class of catastrophes. A number of risks could cause widespread electrical failure, including a series of high-altitude electromagnetic pulses (HEMPs) caused by nuclear weapons, an extreme solar storm, and a coordinated cyber attack. Since modern industry depends on electricity, it is likely there would be a collapse of the functioning of industry and machines in these scenarios. As our current high agricultural productivity depends on industry (for example, for fertilizers) there would be mass starvation in these scenarios with our current understanding. However, there are solutions to our food and nonfood problems in these scenarios, and we plan to do a cost effectiveness analysis of these interventions.
Another project underway relates to single cell protein grown on natural gas. Natural gas is already being used to grow single cell protein for fish. People already eat single cell protein in the form of the supplement spirulina. In a catastrophe, people could eat the single cell protein grown on natural gas. We are investigating whether this would work at the household level. If it does, the waste heat would help heat the person’s house.
We have successfully secured funding to cover operation costs at the current level for 2019. However, there are unrealized acceleration opportunities which additional funding would enable, such as: converting our database of contacts into a self updating useful tool for collaborators, further building the Alliance, and the research below. More ALLFED funding sooner rather than later is valuable because global agricultural catastrophes could occur at any time. Each day acceleration of full preparedness for alternate foods could save 100-40,000 expected lives in the presentation generation, and could increase the value of the far future by 0.000002%-0.002%.
The big picture is that research, development, and planning for alternate foods can be done by people with transferable skills. For instance, experts in biofuels could figure out how to retrofit factories quickly to food production. Therefore, we have great capacity to scale up impact very quickly.
In the next 12 months, we have detailed plans for utilizing $1.5 million, including commitments of people who will work with us if we secure funding. Depending on how the pending proposals fare, the funding gap is between $0.5 million and $1.5 million. Below we outline the projects and the approximate cost.
$20,000 level
Some alternate food technologies that have already been demonstrated, like getting food from wheat leaves. We’d like to produce how-to videos for the techniques and technologies that would be most useful at the household/community scale.
OR
We’ve already begun building a network of existential risk experts. We’d like to continue growing that network, give them media training, and build relationships with the media. Then in a catastrophe, the media would know to contact the panel, and the panel could work to prevent panic and conflict by showing how we could feed everyone.
OR
We have calculated that it is technically feasible to feed all people at least two times over. This means that alternate foods could keep all humans alive and also preserve many other species. We’d like to quantify the cost per expected species saved by alternate foods. These catastrophes could cause extinctions directly, but also starving humans would likely eat other species to extinction. So alternate foods could be a highly effective environmental intervention.
$50,000 level
There is a huge reservoir of fish deep in the ocean. We would estimate potential fish production and cost of scaling up quickly including retrofitting ships into fishing vessels.
OR
Nuclear winter is one of our top concerns. Working with the nuclear winter team funded by Open Philanthropy Project to incorporate alternate foods would help us and them get a better idea of the impacts of nuclear war and how they could be mitigated. This project is analyzing societal impacts and recovery, so the possibility of alternate foods could change the result dramatically.
OR
Estimating the cost and speed of retrofitting existing chemical plants for food production would be valuable. This would likely be in consultation with experts. This could be compared to the option of rapidly constructing dedicated food production facilities.
OR
Initial scoping analysis of new alternate food possibilities, such as single celled protein powered by electricity, direct chemical synthesis of food, seaweed (if the sun is not completely blocked), single celled protein growing on plastic, and mushrooms/Quorn growing on coal/oil/peat.
$100,000 level
Most likely, not all the sun would be blocked by nuclear winter, so it would be very useful to know if we could relocate crops to warmer places. $100,000 would purchase a plant growth chamber to simulate the conditions in the tropics in nuclear winter and also cover the supplies and operator.
OR
Many lives could be saved by planning out how infrastructure could be repurposed for alternate foods, but there is currently little incentive for private organizations to do so. We’d like to create financial mechanisms to incentivise industry preparedness for a significant crop shock, to ensure rapid response in repurposing manufacturing to ensure a sustainable food supply.
Some financial mechanisms we could investigate are catastrophe bonds, parametric insurance and special purpose vehicles and go to industry/governments to pilot financial products that insure food supply whilst creating funds to pay manufacturers to repurpose facilities in a catastrophe. Insurance could be developed that could be paid by governments to fund preparedness.
Or speculatively, perhaps private donors would be interested in funding preparedness and making an agreement with the government. Developed country governments would likely pay exorbitant amounts to feed their citizens in a catastrophe with only stored food. The agreement could be that if this preparedness saved the government $1 trillion or so then the donors would be paid a proportion of savings. If these donors were EAs, they could use that payout to put towards other effective causes. The expected return on investment could be quite high. For instance, from the US perspective, I estimated 800% to 40,000,000% if one were paid the full life-saving benefits (the reduced food expenditure benefits may be similar). We are particularly interested in feedback on this speculative idea.
$300,000 level
GIS analysis of the resources (biomass, industry, etc.) to produce alternate food by country would help with both individual country and cooperative planning. Economic and trade analysis could estimate levels of cooperation at different intensities of food shocks and levels of preparation for alternate foods. This would involve bringing in subject matter experts that we already have agreements with. We would also recruit a graduate student.
$1 million level
Flexible biorefinery investigation: there are factories now that turn crop leaves/stalks into ethanol. They produce sugar first so we could produce human food in a catastrophe. We would perform experiments to produce the leaf protein concentrate, sugar, and feed for chickens, cows, and mushrooms. We have equipment to analyze the nutrients and toxicity. Then we would work out how the process could be scaled up quickly in a catastrophe. This would involve bringing in subject matter experts that we already have agreements with. We would also recruit a significant number of graduate students (which Joshua Pearce at Michigan Technological University and I at University of Alaska Fairbanks have the capacity to manage), so we can scale up quickly. This would be a multi-year project.
How to help
We are always open to feedback and mentoring.
We would appreciate volunteer help on number of projects, including drafting response plans for particular countries (maybe a hackathon?), making alternate foods and documenting instructions and videos, social media (preparation for catastrophe response), etc.
A tax-free donation in the US is easy on our website if you are not donating for the Facebook match.
For the Facebook/Paypal Giving Tuesday matching event on 27 November 2018 for ALLFED, please see here for instructions and here to donate.
Strong-upvoted for presenting a detailed writeup, putting all your work in one place, and including a "Mistakes" section (though I wish you'd written a bit more about how you knew the fundraiser wasn't a good fit after just one month).
Thoughts:
1. What feedback did you get after presenting at DARPA? How was the experience of talking to government officials whose job it is to think about long-term concerns? Did they seem to take the ideas seriously?
(My biggest question about ALLFED has always been "what's the path to reaching policymakers?", and I'm pleasantly surprised to hear about this presentation.)
2. When you say that "It turns out it is much easier to keep most mammal species alive than to feed all people", I can imagine this idea being interpreted as "we'll have to sacrifice some lives in order to preserve biodiversity", which is a message you'll want to be careful about.
At the very least, it seems bad to use language that implies a direct trade-off. For example, you could say something like "it's important to consider how a disaster could affect other species; preventing extinction among edible species will be highly valuable after humanity recovers". (I don't know how close to ALLFED's thinking this is, since I haven't read the book.)
Not that EA doesn't contain plenty of trade-offs like this, of course, but those other trade-offs (e.g. "how many chickens would you rather save than one human?") generally don't appear in scientific research papers. The need to collaborate with governments also seems to weigh against ALLFED's presenting ideas in controversial ways.
Thank you for the feedback and valuable points.
As for how we knew the fundraiser wasn’t a good fit, two factors weighted in our mutual decision to part ways. Firstly, approaching mainstream funders with concerns of existential risk was proving somewhat challenging (this is also the experience that other EA-aligned organizations were telling us they had). Secondly, the fundraiser was finding remote working difficult and discovered he preferred a job with more face-to-face contact. We have a small but intercontinental team, based in the US, Europe and intermittently in Asia, and so our mode of operations at this time is based on networked individuals or teams of two. We have thus discovered we need to highlight this more in our recruitment so as to ensure a good organizational fit for any future hires.
Thanks Aaron!
"what's the path to reaching policymakers?"
I like this question, and obviously we don't want a Yellow Brick Road, leading to a grand figure with no practical power.
[A quick reply, without references, so please read as illustrative/descriptive rather than definitive. This isn't a full answer to your question Aaron, but I hope it gives some impressions of how things are developing in our approach.]
Initially we were keen on reaching policymakers, and we still are to a considerable degree, but we've discovered the following:
So for these and more reasons ALLFED is working:
Nevertheless, government will almost certainly be important in an event, and getting GCR/X into disaster preparedness (aka "mainstreaming") would be great.
One way to do this is via NASA into the UN's Sendai process, and I will working on text for that in
later this year.
Overall, policy people do often rely on the seniority of scientists to tell them who to choose to listen to about which risks to take on and how. (Obviously, seniority is not necessarily the best criteria, especially with emerging tech! But that's the realpolitik, and what we have to work with.) So ALLFED needs a bigger repertoire, more heavyweight policy institutes backing us, and a wider network of academics who have "bought in" to our line on cost effectiveness and duty-of-care (for nations to protect their population).
Academics, like everyone else, can be conservative and scared of ridicule, so ALLFED is emphasising GCR more than X-risk, not because we think X-risk is less important, but because in order to reach policymakers you have to
(a) be able to communicate about things that they can conceive of and grapple with
(b) help them not fear attack in the press for being too sci-fi
(c) give them clear "realistic" justifications for their own finance people.
And again, the volcanic GCR example is very helpful here because we have a clear historical precedent or two (Tambora and Laki) that politicians can relate to. It's also easy to convince them that another VEI7 (or worse) is certain to happen some day.
We are also working on fall back strategies in case there is a GCR/X-risk event in the next year or two, so that ALLFED is of immediate and practical usefulness to some governments/media and industry.
One area it would be great to have specific funding for, as a self-contained project, is a self-updating GCR/X expert directory. Almost everyone we network with wants one of these, and no one has cash/staff to do it. I'd like to see an India EA project funded to do it.