Concrete research projects
The version of the browser you are using is no longer supported. Please upgrade to a supported browser.Dismiss

View only
Project NameCause Areastype of workestimated timeDescriptionSkills and background neededPeople and organizations who have expressed interestChances of success further readingcommentsWho's taken up this project?
Explanation of this table:Click this for the guide!
For errors or suggestions: Edo at
Literature review of Privacy Preserving
Future of Civilization; Governance; Emerging TechnologiesLiterature Review4-10 DaysThere is a debate on whether we should have more or less surveillance.
On the one hand, more surveillance may help with future terrorism mitigation (particularly if terrorists will have much easier way of generating weapons of mass destruction).
On the other hand, it increases the possibility of a capable authoritarian government.
In an EAG presentation, Ben Garfinkel proposes that we can help both sides of the argument by improving the possibilities of preserving privacy while using mass surveillance.
The first step here is to collect all relevant discussions and literature and write up a forum post.
Familiarity with modern cryptographic tools and relevant technologies is helpful but not necessary. Familiarity with the EA conversation on longtermism. Ben Garfinkel from FHIHighThe Future Of Surveillance - Ben Garfinkel EAG TalkPerhaps related
Tom Kalvary
Literature review of historical effects of moral philosophy History, Moral Circle Expansion, Moral PhilosophyLiterature Review2-6 Days
In the above post, the authors conducted a quick research on the topic of the impact of moral philosophy on major moral shifts in society. Specifically, they looked at the the cases of women's rights and animal welfare.

In the end, they asked "What previous work exists examining the historical role of moral philosophy?". A first step here is to find and briefly explain the conclusions of such works.
Familiarity with historical analysis can help.
Alex Hill and Jaime SevillaHighThe post and it's comments
What aging-biomarker methods might be best for measuring the cumulative welfare of juveniles versus adults?Animal WelfareLiterature Review5-40 DaysIn this post:
Will Bradshaw of WAI outlined the use of biomarkers of aging as a proxy for stress and thus welfare of animals.

In a following post
he outlines many directions for continued research. For generalists without a lab or expertise- given some reading up on evolutionary biology and comparative psychology, he thinks progress could be made on many of the questions in section 2 without access to a laboratory, both through focused literature reviews and greater theoretical scrutiny.

One possible project here is to collect the relevant literature and give a first take on what methods might be best for measuring the cumulative welfare of juveniles versus adults?
basic knowledge of evolutionary biology and comparative psychology.Wild Animal InitiativeMedium-High
How many bees are transported world-wide?Animal WelfareLiterature Review, Estimation1-3 DaysJason Schukraft from Rethink Priorities wrote a lengthy report on Honey Bee Welfare.

In the last section, there are many further research questions. The first one is about the effects of transport on bees.
Long-distance transport has worrying effects on honey bee physiology. Ahn et al. 2012 found that “bees experiencing transportation have trouble fully developing their food glands and this might affect their ability to nurse the next generation of workers”.

How many bees are transported worldwide? What is the average length of transport? What are the transport conditions like outside the U.S.?
Estimation. Finding information.Rethink PrioritiesMedium-High
Brainstorm a list of possible policies we may want NGOs to advocate forPolicy, AdvocacyLiterature review, Interviews, Surveys1-5 DaysIn a report by Charity Evaluator about Animal Research,,
they analyzed and compared many different possible new research organizations ideas.

On the top of their list they have "Animal Ask Research", which tries to figure out what are the best policies we should advocate for ("Asks"). Like pretty much everything in this list, a proper answer for that requires a lot of work and hence the call for a new organisation. However, it is definitely possible to make headway and have original contribution.

What Asks should someone advocate for? Pick a cause area (Doesn't have to be Animal Welfare), speak with relevant people to get a sense of the situation, brainstorm what Asks should be advocated for.
Communication. Familiarity with a given area, or people there. Charity Evaluator.Low-MediumNote that many Asks are already pursued. In the case of animal advocacy, one should compare the list with data from Faunalytics and Animal Charity Evaluator
Analyze Electoral Data from Australian Federal ElectionsImproving Institutional Decision MakingData Science Exploratory research5-20 DaysThe Australian government publishes electoral data. Since they use “instant-runoff voting” it may be very interesting to mine for information on problems with their voting system.
Understanding of the science of voting systems, or the prerequisites to learn that. Basic data analysis skillsThe Center for Election ScienceLow-Medium'm not sure, but it may be possible to answer this post on the CES forum -

There seems to be an active discussion in the CES forum on research related to voting theory. There's bound to be a large amount of possible research questions that are directly interesting. It is aimed at experts.
Review considerations regarding the chance of local, fast takeoff AGIAI SafetyCritical Review, Argument analysis 1-5 DaysAI Safety work is critically dependent on estimates of the time, duration and magnitude of the creation of Artificial General Intelligence (AGI). In particular, a local and fast takeoff (in which, say, a single innovation makes creating AGI easy to immediately deploy) is particularly risky.

Collecting arguments on this topic is helpful because opinion among well-informed thinkers on the topic seems to diverge from what would be expected given the arguments that we know about. This suggests that we are missing some important considerations, that we would need to well assess the chance of local, fast takeoff.

AI Impacts, an organisation devoted to analysis of estimated timeline for AGI, have a list of considerations for that case. If you find local, fast take-off likely, check if the considerations that lead you to this view are represented. Alternately, interview someone else with a strong position about the considerations they find important. If there are any arguments or counterarguments that you think are missing, write a short page explaining the case.
Familiarity with the existing discussion is needed.
AI ImpactsLow-MediumThe AI Impacts article that collects all evidence and reasons they have to either side of the debateAI Impacts has a large list of possible research questions -

for some more ideas about how to contribute to AGI Safety without technical skills in AI
Disentangling a debate regarding AI Failure Scenarios between Robin Hanson and Paul ChristianoAI SafetyArgument Analysis1-2 DaysIn
Paul Christiano explains how he views the likely catastrophe that may result from failure of AI Alignment.

Robin Hanson makes a response in which he claims that this scenario results from a failure of solving a Principle-Agent problem which seems similar to other cases studyed in Economics.

There seems to be a mismatch between what they are saying that has not been thoroughly addressed. The next step here is to publish a post on LessWrong with a short analysis of both sides of the "debate", and try to clarify what are the important disagreements.
Basic understanding of concepts in economics and AI alignment. Familiarity with the context of the discussion. A researcher from FHIMedium-Highseems related:
Estimating how important is algorithmic progress to AGI timelineAI SafetyProgramming, AI5-30 DaysHow much does algorithmic progress changes over time? This is clearly important for AI timeline estimation, but how can we quantify it? Using standard measures for AI progress conflates with it also progress in hardware.
One (fun) option is to see how state-of-the-art algorithms run on old hardware, and measure the results as a function of hardware.
AI and Programming expertise is needed.A researcher from FHIMedium
Summarise a paper that analyzes market design solutions to donor coordinationsAltruistic CoordinationAnalysis of a technical paper.1-3 DaysSection 2.7 In GPI's research agenda,
is about Altruistic Coordination. That is, how can different people that want to do good can collaboratively do better.

A specific question there is how can donors coordinate their contributions so that there are no perverse incentives that results from a limited room for funding. A possible first step here is to write a summary of a paper on the subject with some impossibility results, and write up the main relevant takeaways:
Basic knowledge of market design, economics and computational complexityGPIHighDominik Peters -
Critically review the Symmetry Theory of ValenceConsciousness researchArgument Analysis3-10 DaysThe Qualia Research Institute studies consciousness with the goal of using the results to have better interventions and ways of improving the world.
They are basing their research on the hypothesis that valence is associated with symmetry.

In a recent comment,,
Mike Johnson (the founder of QRI) says that they are looking for a capable researcher to provide a critical review of their work. Since this topic is somewhat fuzzy and there is a variety of existing expertise and discussions around similar ideas, it may be hard to provide a useful criticism without the necessary background.

I think that in a short period of time, it may be possible to check whether the predictions made here:,
can be more concrete and how likely are the results of the experiment to say something meaningful about the hypothesis.
basic familiarity with neuroscience, consciousness and mathQualia Research InstituteLow-MediumShould probably first look through the Pingbacks of that page
Evaluate a report by 80K on what are the most important skills to learnMeta, Career adviceData Analysis1-5 DaysAs EA is rather new, it is important to evaluate existing research.

This report,
attempts to find what are the most useful skills to learn, and reaches some unintuitive conclusions written for general audience here-

There is a clear caveat in the 80K article that the methodology is rough and has weaknesses, and there are many questions for further research in the technical report itself. A first step here is to verify the analysis, and perhaps run some more tests.
Data analysis80KHigh
Collect all possible Task-Y candidates (Scalable part-time tasks which people interested in EA can do to help)Community Building, MetaLIterature Review1-2 DaysIn this forum post:,
Alexrjl presents a need for figuring out what projects we can give part time volunteers that can scale at large.

Collect relevant ideas from forum posts and comments, and search online for other ideas and some implementations of possible Task-Ys. Write it all down in a forum post.
Familiarity with the EA communityHighShould probably first look through the Pingbacks of that page
Expend on a Wikipedia article on Audience Cost - The cost for a country leader from backing down in an escalated conflictGlobal PeaceWriting1-2 DaysBrian Tomasik has a list of topics on wikipedia he'd like to see updated:
(Note that it hasn't been updated since 2016).

Among the suggestions there, there is the short Wiki page on Audience cost:

Expand on it, explain why it is important, and link to that page from other relevant sources.
International relationsBrian TomasikHigh
Summarise Chapter 2 of Nick Beckstead's dissertation On the Overwhelming Importance of Shaping the Far FutureLongtermism, Global Priorities ResearchPaper Summary2-4 DaysNick's dissertation,,
is an important paper regarding our ethical considerations for the long term future.
An explanatory post aimed at general EA audience that tries to explain some concepts of it can be useful.
Chapter 2 also engages with some psychological claims, so I think that may be a good starting point.

Write a (short) forum post that explains the key ideas from chapter 2 (or perhaps just 2.1, 2.2).
Familiarity with the subjectA researcher from GPI said that it might be a good and valuable project to summarize some of the basic papers in the field to non-experts.High
Contribute to the LessWrong and EA forum codeCommunity Programming, Web Development1-3 DaysThe EA forum is forked from LessWrong. JPAddison is the person at CEA in charge of the EA forum Fork. Most of the work is thus done on LessWrong.

Tackle an open issue at
Programming, WebHighFrom a first view, the code is not very open-source friendly (the tags do not seem to be maintained, for example).
Predicting the results of the coming EA SurveyMetaForecasting1-4 is a platform for collaborative forecasting. In the following link are some questions related to the coming EA survey.
These forecasts are an experiment in amplifying researchers and decision makers, so this should be thought of as a prototype for possible future infrastructure.

Familiarize yourself with basic forecasting methods, and tackle a bunch of these questions.
Before doing the forecasting, should be calibrated and familiar with Forecasting in general.High
Make sure that Wikipedia has good pages for basic cognitive tools from ConceptuallyImproving RationalityWriting1-10 is a site founded by EAs that sends a weekly newsletter with a new concept each week that is meant to serve as a new cognitive tool.

The concepts there were carefully selected, and written to be clear and concise. Go through each of these concepts, and make sure that their wikipedia page exists and is on a high quality.
Familiarity with the basic cognitive tools is usefulWill MacAskillHigh
Collect Writeups about Institutional Change and GovernanceInstitutional ChangeLiterature Review1-3 DaysWhile outside of EA there is a lot of discussion about governance models, aligning incentives and systemic change, it is not organized well in EA. The conversations in EA on these matters seems scattered and I (Edo) do not see it as a coherent "cause area" of this semantic neighborhood, even though there seems to be a strong common core. There is work related to improving institutional decision making, GPI has some works on future-generations-oriented government, there is work on voting methods by the Center For Election Science, there is work by Nick Bostrom on World-Governments, there are also forecasting, infrastructural and methodological works on better decision making processes.

As a first step, collect many different aspects of "governance" to a single forum post. You can try to link them together or create a taxonomy.
Familiarity with the concepts involvedHighThere are some discussions of systemic change in EA ( I mostly found them to be about "changing the system" rather than "creating a system", which may or may not be more important but we should be clear of the distinction