Quite wonderfully, there has been a proliferation of research questions EAs have identified as potentially worth pursuing, and now even a proliferation of collections of such questions. So like a good little EA, I’ve gone meta: this post is a collection of all such collections I’m aware of. I hope this can serve as a central directory to all of those other useful resources, and thereby help interested EAs find questions they can investigate to help inform our whole community’s efforts to do good better.
Please also leave a comment if you know of any research question collections that aren't yet included in this post!
Some things to note:
- It may be best to just engage with one set of questions that are relevant to your skills, interests, or plans, and ignore the rest of this post.
- It’s possible that some of these questions are no longer “open”.
- I’ve included some things that aren’t explicitly written as collections of research questions, as long as research questions could very easily be inferred from them (e.g., from the problems people identify, or the posts people want written).
Most recent update to this post (typically adding a link to another collection): December 2022.
Various EA-related topics
- List of EA-related thesis topics - Effective Thesis, no date
- You can also contact Effective Thesis for discussion, help, or coaching.
- Why EA needs Operations Research: the science of decision making - wesg, 2022
- Some research questions that you may want to tackle - Aaron Bergman, 2022
- On the longtermist case for working on farmed animals [Uncertainties & research ideas] - Michael Aird, 2021
- What are the highest impact questions in the behavioral sciences? - Abby Hoskin, 2021
- What are the key ongoing debates in EA? - various, 2020
- What posts do you want someone to write? - various, 2020
- What questions could COVID-19 provide evidence on that would help guide future EA decisions? - Michael Aird (i.e., me) and others, 2020
- Some history topics it might be very valuable to investigate - Michael Aird, 2020
- A collection of researchy projects for Aspiring EAs - EdoArad, 2019
- What are some lists of open questions in effective altruism? - Aaron Gertler and others, 2019
- This and the following list are roughly the same sort of “meta collection” as this post, and I think I took everything relevant from them already.
- Technical and Philosophical Questions That Might Affect Our Grantmaking - Open Philanthropy Project, 2017
- 2018 list of half-baked volunteer research ideas and its comments - Jacy Reese and others, 2018
- EA Summit Project Ideas (specifically the “Research Projects”) - various, no date
- The most important questions and problems - Pablo Stafforini
- EA Research to be done in Philosophy - I don't know who created this list
Mostly focused on longtermism, existential risks, or GCRs
- Psychology for Effectively Improving the Future — A Research Agenda - Effective Altruism Psychology Lab, 2022
- Nuclear risk research ideas - Michael Aird (me), 2022
- Legal Priorities Research: A Research Agenda - Legal Priorities Project, 2021
- Project Ideas in Biosecurity for EAs - David Manheim ("In conjunction with a group of other EA biosecurity folk"), 2021
- Politics, Policy, and Security from a Broad Longtermist Perspective: A Preliminary Research Agenda - Michael Aird for Rethink Priorities, 2021
- Humanities Research Ideas for Longtermists - Lizka for Rethink Priorities, 2021
- 80 Questions for UK Biological Security - Luke Kemp et al., 2021
- The Precipice, Appendix F: Policy and research recommendations - Toby Ord, 2020
- Research questions that could have a big social impact, organised by discipline - Arden Koehler & Howie Lempel (80,000 Hours), 2020
- Crucial questions for longtermists - Michael Aird for Convergence Analysis, 2020
- Some history topics it might be very valuable to investigate - Michael Aird, 2020
- Questions related to moral circles that are listed at the end of this post and in this comment - Michael Aird, 2020
- Cause prioritisation / macrostrategy topics Denis Drescher collected and may investigate - 2020
- Michael Aird_Research statement [FHI RSP] - Michael Aird, 2020
- Open Research Questions - Center on Long-Term Risk, no date
- ALLFED’s research priorities and Effective Theses topic ideas - 2019
- Open Research Questions - Center for Reducing Suffering, no date
Mostly focused on AI
Technical/theoretical AI safety/alignment
- Some conceptual alignment research projects - Richard Ngo, 2022
- AGI Safety Fundamentals' list of AI alignment project suggestions - 2021
- AI Safety Research Project Ideas - Evans and Armstrong, 2021
- “clusters of ideas that we believe warrant further attention and research” - Center for Human-Compatible AI (CHAI), no date
- Research Agenda v0.9: Synthesising a human's preferences into a utility function - Stuart Armstrong, 2019
- Related talk here
- The Learning-Theoretic AI Alignment Research Agenda - Vanessa Kosoy, 2019
- FLI AI Safety Research Landscape - Future of Life Institute, 2018
- Associated paper here
- Agent Foundations for Aligning Machine Intelligence with Human Interests: A Technical Research Agenda - Nate Soares and Benya Fallenstein (MIRI), published originally in 2014, and in 2017 in The Technological Singularity
- Alignment for Advanced Machine Learning Systems - Jessica Taylor et al. (MIRI), 2016
- Concrete Problems in AI Safety - Amodei et al., 2016
- Some things by or related to Paul Christiano that some people have indicated serve as research agendas, collections of questions, or supporting materials:
- Iterated Distillation and Amplification
- Paul's research agenda FAQ
- AI alignment landscape
- Directions and desiderata for AI alignment
- Note: I haven’t checked most of these out myself
- There may be other research agendas listed here
AI policy/strategy/governance
- Questions for further investigation of AI diffusion - Cottier, 2022
- AI Forecasting Research Ideas - Sevilla, Heim, Ho, 2022
- Some AI Governance Research Ideas - the Centre for the Governance of AI, 2021
- AGI Safety Fundamentals' list of AI governance project suggestions - 2021
- Problems in AI risk that economists could potentially contribute to - Michael Aird, 2021
- (AI × social science) Research Questions - Vael Gates, 2021
- Open Problems in Cooperative AI - Dafoe et al., 2020
- Forecasting AI Progress: A Research Agenda - Gruetzemacher, Dorner, Bernaola-Alvarez, Giattino, & David Manheim, 2020 (comments here)
- Longtermist AI policy projects for economists - Risto Uuk (this doc was originally just made for Risto's own use, so the ideas shouldn't be taken as high-confidence recommendations to anyone else)
- Cooperation, Conflict, and Transformative Artificial Intelligence (the Center on Long-Term Risk’s research agenda) - Jesse Clifton, 2019
- Problems in AI Alignment that philosophers could potentially contribute to - Wei Dai, 2019
- Technical AGI safety research outside AI - Richard Ngo, 2019
- A shift in arguments for AI risk - Tom Sittler, 2019
- The Centre for the Governance of AI’s research agenda - 2018
- Promising research projects - AI Impacts, 2018
- They also made a list in 2015; I haven’t checked how much they overlap
- Artificial Intelligence and Global Security Initiative Research Agenda - Centre for a New American Security, 2017
- A survey of research questions for robust and beneficial AI - Future of Life Institute, no date
- Cognitive Science/Psychology As a Neglected Approach to AI Safety - Kaj Sotala, 2017
- “studies which could illuminate our strategic situation with regard to superintelligence” - Luke Muehlhauser, 2014 (he also made a list in 2012)
A mix of technical and non-technical
- Important, actionable research questions for the most important century - Holden Karnofsky, 2022
Mostly focused on biorisk or coronavirus
- Coronavirus Research Ideas for EAs - Peter Wildeford, 2020
- LessWrong Coronavirus Agenda - Elizabeth, 2020
Cause prioritisation/global priorities
- The Global Priorities Institute’s research agenda - 2019
- The most important unsolved problems in ethics - Will MacAskill, 2012
- I’m guessing this is mostly superseded by GPI’s agenda, but I haven’t checked.
Animal welfare
- Alternative Protein Solutions Database - Good Food Institute, 2020
- List of "burning questions" - Tiny Beam Fund, 2020 (see here for some more context) (note: I don't know much about Tiny Beam Fund or how EA-like their thinking and priorities are)
- Animal Advocacy Research Fund’s Focus Areas - Animal Charity Evaluators, no date
- Wild Animal Initiative’s research agenda - 2019
- Sentience Institute’s research agenda - 2019
- Sentience Institute's summary of "Foundational Questions for Effective Animal Advocacy" - 2019
- “Less explored” foundational questions in effective animal advocacy - Sentience Institute, 2019
- Faunalytics’s Research Priorities - 2019
- Alternative Proteins: 2020 Consumer Research Priorities - Good Food Institute, 2019
(Perhaps Charity Entrepreneurship and Rethink Priorities have relevant collections or research agendas?)
Global health and development
Important unresolved research questions relevant to macroeconomic policy - Open Philanthropy Project, 2014
(Perhaps Charity Entrepreneurship and GiveWell have relevant collections or research agendas?)
Other areas many EAs are interested in
Forecasting & improving institutional decision-making
- Some research ideas in forecasting - Sevilla, 2022
- How valuable would more academic research on forecasting be? What questions should be researched? - Michael Aird, 2020
- Research Directions on Improving Policymaking - EA Geneva, probably 2020
Rationality
- What are the open problems in Human Rationality? - Raemon, 2019
- Note: I haven’t read this and don’t know how well it fits here.
Mental health, happiness, etc.
- Happier Lives Institute’s research agenda - Michael Plant, 2019
- Qualia Research Institute’s research agenda, no date
- Health and happiness: some open research topics - Derek Foster, 2019
Other?
I’d guess there are other relevant areas for which research questions have been collected somewhere.
Potential lists of lists which I haven’t properly taken the lists from yet
- What are EA project ideas you have? - Mati_Roy and others, 2020
- What new EA project or org would you like to see created in the next 3 years? - Ozzie Gooen and others, 2019
- Concrete project lists - Richard Batty, 2017
Thanks to all the people who created all the lists I’ve shown and/or taken from here. And thanks to Aaron Gertler for his above-quoted thoughts, to David Kristoffersson for helpful feedback and additions, and to Remmelt Ellen for helpful comments.
This post is related to my work with Convergence Analysis.
This was a very practical post. I return to it from time to time to guide my thinking on what to research next. I suggest it to people to consider. I think about ways to build on the work and develop a database. I think that it may have helped to catalyse a lot of good outcomes.
Not central to EA, but there's Gwerns open questions.
For calibration, so far no one has contacted me to take on one of the research projects in the list of concrete researchy projects. And even in 1-1s, with people that are interested in joining EA Israel and are interested in taking on research project, it had very limited success going over this list and thinking together about possible research questions.
(Upvoted)
Yeah, I've seen that sort of thing mentioned a few times, such that I no longer find it surprising, though I initially did, and I still don't fully understand why it's the case.*
That's why I included "I think we could do more to inspire and support people to actually investigate these questions than just assemble a big list", and the points after that. But I'd definitely be keen to hear more thoughts on how to provide effective inspiration and support for that. (Indeed, it seems that could be a research question in itself. Now, if only we could inspire and support people to investigate it...)
*It does seem there are a lot of interesting and important questions to be explored, many of which may not require extremely specialised skills. As well as a lot of intellectually curious, research-minded EAs interested in having more EA-y things to do. So my guess before hearing that sort of thing mentioned a few times probably would've been that there'd be more uptake of these sorts of lists, and I'm not entirely sure what ingredients are missing.
Obviously payment and organisational infrastructures would be very helpful for most people, and necessary for many. But I wouldn't guess they'd be necessary for all curious EAs with some slices of free time? I wonder if there are other levers that could be pulled to unlock some of this extra talent that seems to be floating around?
My current model is something like this. #BetterWrongThanVague
It is difficult to make noticeable research contribution. Even small incremental steps can be intimidating and time consuming.
It is hard to motivate oneself to work alone on someone else's problems. I think that most people probably have their own passions and model of what's important, and it's unclear why subquestion 3.5.1 should be the single thing that they focus on.
Three of the main motivators that might mitigate that here are recognition for completing the work well and presenting something interesting, better career capital (learning something new or displaying skills) and socializing/partnering.
One thing which I thought about trying which might be related is to take on a small scale research problem and set up an open call to globally collaborate on this. To make it successful, we can set up something formal that some organisation is interested in this result (and better yet, possibly supply a prize - doesn't have to be monetary) and coordinate with local groups to collect an initial team.
That could be fun and engaging, but I'm not sure how scalable this is and how much impact we can expect from that (which is uncertainty probably worth of testing out). I've tried to start a small ALLFED-directed research group locally, as part of our research team, but that also didn't work out. I think that going global might possibly work though.
Hey, thanks for putting this together. I think it would be quite valuable to have these lists be put up on Effective Thesis's research agenda page. My reasoning for this is that Effective Thesis's research agenda page probably has more viewers than this EA Forum post or the Google Doc version of this post.
Additionally, if you agree with the above, I'd be curious to hear your thoughts on how we could make Effective Thesis's research agenda page open source?
I think those are both good ideas! (This is assuming that by "open source" you mean something like "easy for anyone to make suggestions to, in a way that lets the page be efficiently expanded and updated". Did you have something else in mind?)
I don't know the Effective Thesis people personally (though what they're doing seems really valuable to me). But I've now contacted them via their website, with a message quoting your comment and asking for their thoughts.
Yep, that's what I meant by "open source"! Awesome to hear you're taking this forward!
Update: Effective Thesis have now basically done both of the things you suggested (you can see the changes here). So thanks for the suggestions!
Glad to hear this!
Noticeable lack of Global Health and Development lists/topics, particularly as this is where most of the individual giving where most of the individual EA giving is going. Hope I can help with this at some point.
I still think this was a useful post. It's one of many posts of mine that seem like they were somewhat obvious low-hanging fruit that other people could've plucked too; more thoughts on that in other self-reviews here and here.
That said, I also think that, at least as far as I'm aware, this post has been less impactful and less used than I'd have guessed. I'm not actually aware of any instances where I know for sure that someone used this post to pick a research project and then followed it through to completion. I am aware of two other well-received and high-quality-seeming research question collection posts that were substantially influenced by this one (by 80k and by Lizka), but I'm also not aware of clear signs of impact from those either. This seems kinda surprising and sad.
I wrote some thoughts on why that might've happened, and what it might suggest we need to do to improve the EA-aligned research pipeline, in this sequence and in a draft "Proposal: A central, editable database to help people choose and do research projects". But I no longer plan to publish the latter draft or to strongly encourage people to work on that idea. (I can expand on why if people are interested.) I'm now focused more on other ways to improve the EA-aligned research pipeline, like causing there to be more available high-quality mentorship via research training programs, scaling EA research orgs, and sharing tips and resources on how to do research and how to do management and stuff.
Thanks so much for this! I am keen to discuss this when Covid-19 has passed. I have some ideas and see opportunities for collaboration. EdoArad - I would love to talk with you too at that time. For context, I am one of the people involved in READI which is led by EA volunteers and seeking to tackle high impact/EA aligned research questions. This is our current project, you can see other work here.
I'm very interested in the work you are doing at READI, and it would be great to discuss ideas and collaborate.
(by the way, what does READI stand for?)
Here's a section that was in this post until today but that I've now decided is no longer worth having in the post itself:
What this could become (with your help!)
[Edited to add in April 2021: I've now drafted a post that goes into more detail on a better version of the sorts of ideas given below.]
As noted earlier, I hope this can help some of the many wonderfully curious EAs out there to find important questions they can start plugging away at, to help guide us all in our various efforts to improve the world.
But I’m sure that:
And I think we could do more to inspire and support people to actually investigate these questions than just assemble a big list. For example, we could somehow “attach” to each question, perhaps as comments or indented bullet points, things like:
And this could all be done collaboratively. (Plus, I don’t expect to have time to do it myself.)
So here’s a Google Doc version of this post. [Edit: I'm no longer updating that when I update the post version.] Anyone can comment and make suggestions. Please do so, to make this as useful as it can be! (You can either say “someone should probably do X”, or just do go do X yourself.)
Also feel free to:
I’d be really excited to see this develop into something that can really help people advance our movement’s collective knowledge, and to see people actually executing on that - actually making those advancements.
Thoughts from Aaron Gertler
I emailed Aaron Gertler of the Centre for Effective Altruism to ask his thoughts on how valuable something like this would be, and what its ideal eventual form might be. His reply, which he confirmed it was ok for me to quote here, included the following:
Hopefully that can provide food for thought for people who might want to develop this idea further.
I would add the GFI alternative protein solutions database
Thanks! Added.
Addition: Open Problems in Cooperative AI.
Thanks! Added.
Additions under "Less technical / AI strategy / AI governance"?
- https://forum.effectivealtruism.org/posts/WdMnmmqqiP5zCtSfv/cognitive-science-psychology-as-a-neglected-approach-to-ai
- https://forum.effectivealtruism.org/posts/9kNqYzEAYtvLg2BbR/baobao-zhang-how-social-science-research-can-inform-ai (though this one only has three research questions and isn't focused on generating questions)
Thanks! I've now added the first of those two :)
Got my post up :). https://forum.effectivealtruism.org/posts/dKgWZ8GMNkXfRwjqH/seeking-social-science-students-collaborators-interested-in
Also "Artificial Intelligence and Global Security Initiative Research Agenda - Centre for a New American Security, no date" was published in July 2017, according to the embedded pdf in that link!
Thanks for the heads up - I've now added a link to your doc and changed the date for the CNAS agenda :)
Update: 80,000 Hours have released an article entitled Research questions that could have a big social impact, organised by discipline, which draws on the lists of questions listed by this post, but also includes some new questions (sometimes from personal correspondences with the authors). Readers may want to check that article out too. (I've now added a link from this post.)
Note that there is this document of projects and exercises relating to AI safety, which is part of the Stampy project.