In the context of EA, Shapley values are a method for assigning credit for the impact of an intervention to each of a set of actors collaborating to make it happen. The concept of a Shapley value comes from cooperative game theory, where it is a general solution to the problem of distributing gains from cooperation. In impact assessment, it is often compared...
The fragile world hypothesis is the hypothesis that 'if technological development continues indefinitely, systemic fragility will increase to the point that the possibility of a shock sufficient for complete collapse approaches certainty.'[1]
Closely related to the concept of a 'global polycrisis':
...Established concepts, such as “systemic risk” (Renn 2016; Renn
Charity Entrepreneurship's Research Training Program aims to train aspiring researchers to rapidly produce high-quality decision-relevant research.
Ambitious Impact | Research | Cause prioritization | Cause candidates
Far-UVC (also far-UV or far-UV-C) is light with a wavelength of 200–235 nm. It is a type of germicidal ultraviolet (GUV) light, which includes the whole UVC and parts of the UVB spectrum. Far-UVC has emerged as a highly promising approach to indoor air disinfection, offering enhanced effectiveness and safety compared to conventional 254 nm upper-room GUV systems...
The Open Philanthropy AI Worldviews Contest was a 2023 competition organized to surface novel considerations that could influence Open Philanthropy’s views on AI timelines and AI risk. A total of $225,000 in prize money was awarded across six winning entries. The contest served as the formal successor to a 2022 preannouncement and as a spiritual successor to...
The Unjournal is an organisation that works to organize and fund public journal-independent feedback, rating, and evaluation of hosted papers and dynamically-presented research projects. Their initial focus is on quantitative work that informs global priorities, especially in economics, policy, and social science. They aim to encourage better research by making...
Value erosion refers to a process by which competitive dynamics could eventually lead to "the proliferation of forms of life (countries, companies, autonomous AIs) which lock-in bad values", if those forms of life outcompete others (Dafoe, 2020).
Dafoe, Allan (2019) Value erosion for FHI July 2019, July 25.
Bostrom, Nick (2004) The...
When human-level AI will emerge (AI timelines) is a consideration for both AI risk interventions and other interventions. For example, one's AI timelines affect how much to discount interventions that have a delayed effect.
The Asymmetry is a view in population ethics stating that creating bad lives is morally bad, but creating good lives is not morally good. That is, there is an asymmetry between (i) the strong moral reason to avoid creating lives filled with suffering and (ii) the absence (or weakness) of a corresponding reason to create happy lives.
This asymmetry is often used...
This topic is for posts discussing Large Language Models (LLMs) -- for example, the GPT models produced by OpenAI.
AI safety | Artificial intelligence | AI governance | AI forecasting
| User | Post Title | Topic | Pow | When | Vote |
Effective Altruism Community Building Grants (or the CBG programme) (CBG) is a project rungrantmaking program supporting professional city and national effective altruism groups. The program was established in 2018 by the Groups team at the Centre for Effective Altruism (CEA), providing full-.
CBG provides funding to a set of established EA groups. Historically, CEA also provided non-monetary support to grantees, including retreats, regular check-ins, coordination calls, a shared Slack space, and part-timeresources for running groups.
In 2026, CEA began transitioning evaluation of CBG grants to individuals andEA Funds, where the portfolio is expected to be managed alongside, but remain separate from, the EA Infrastructure Fund. CEA also phased out or transitioned most CBG-specific non-monetary support. Funding for existing CBG groups doing local effective altruism community building.[1] It was started in 2018. continues, but the portfolio remains closed to new applicants.
Centre for Effective Altruism | Effective Altruism Funds | Effective Altruism Infrastructure Fund will start evaluating grants for some EA | Effective altruism groups, Joan Gass.
DiGiovanni offers a comprehensive list of potential objections to his argument, and their rebuttals, but heas well as clarifications on what this argument does (not) claim. He also mentions two attempts to find action-guidance despite cluelessness: i) consequentialist bracketing, and ii) a metaepistemic wager.
DiGiovanni, Anthony. 2026.2026a. “Clarifying some misunderstandings of the unawareness argument.” Sept 1. https://forum.effectivealtruism.org/posts/LiicMNNYiLpocw3wZ/clarifying-some-misunderstandings-of-the-unawareness
DiGiovanni, Anthony. 2026b. “Cluelessness: Summary of the argument, why it matters, and counterarguments.” June 19. https://forum.effectivealtruism.org/posts/NesgdEY6yPrE9wDap/cluelessness-summary-of-the-argument-why-it-matters-and-1
Forethought is "a research nonprofit focused on how to navigate the transition to a world with superintelligent AI systems".
Forethought is "a research nonprofit focused on how to navigate the transition to a world with superintelligent AI systems".
Perry, Lucas (2019) AI Alignment Podcast: On the governance of AI with Jade Leung,AI, Future of Life Institute, July 22.
Kagan, Rebecca, Jade Leung[redacted], & Ben Clifford (2021) What we learned from a year incubating longtermist entrepreneurship, Effective Altruism Forum, August 30.
Bergal, Asya (2021) Long-Term Future Fund: May 2021 grant recommendations, Effective Altruism Forum, May 27.
Kagan, Rebecca, Jade Leung[redacted], & Ben Clifford (2021) What we learned from a year incubating longtermist entrepreneurship, Effective Altruism Forum, August 30.
For example, there are some reasons to think that the long-term effects of a marginally higher economic growth rate would be good—overall good or c-preferable—say, via drivingby fostering more patient and pro-social attitudes. This would mean that taking action to increase economic growth could have much better effects than not taking the action. We have some reasons to think that the long-term effects of a marginally higher economic growth rate would be bad —for example, via increased carbon emissions leading to climate change. This would mean that not taking the action that increases economic growth could be a much better idea. It is not immediately obvious that one of these is better than the other, but we also cannot say they have equal expected value. Rather, the EV is indeterminate.
We care about the overall effects of a marginally higher economic growth rate (~P1). If we have no reason to believe that our best guess (on whether it is overall net good or bad) is truth-tracking because of substantial unawareness, we should not follow this best guess (~P2). Given how coarse our awareness of the possible consequences of a marginally higher economic growth rate in fact is, we indeed are in such a situation (~P3).
DiGiovanni offers a comprehensive list of potential objections to his argument, and their rebuttals, but he also mentions two attempts to find action-guidance despite cluelessness: i) consequentialist bracketing, and ii) a metaepistemic wager.
While the term consequentialist bracketing was first used by DiGiovanni (2025b)[3], the idea was formalized by Kollin et al. (2025). It basically aims at providing a principled way of excluding some paralyzing considerations from our decision-making (i.e., "bracketing them out") to find action-guidance, butguidance. Still, it comes with (at least) the following potential problems:
DiGiovanni offers a comprehensive list of potential objections to his argument, and their rebuttals, but he mentions two attempts to find action-guidance despite cluelessness: i) consequentialist bracketing, and ii) a metaepistemic wager. Both are discussed below, right after we get more precise on DiGiovanni's empirical premise.
For example, there are some reasons to think that the long-term effects of a marginally higher economic growth rate would be good—say, via driving more patient and pro-social attitudes. This would mean that taking action to increase economic growth could have much better effects than not taking the action. We have some reasons to think that the long-term effects of a marginally higher economic growth rate would be bad —for example, via increased carbon emissions leading to climate change. This would mean that not taking the action that increases economic growth could be a much better idea. It is not immediately obvious that one of these is better than the other, but we also cannot say they have equal expected value. We care about the overall effects of a marginally higher economic growth rate (~P1). If we have no reason to believe that our best guess (on whether it is overall net good or bad) is truth-tracking because of substantial unawareness, we should not follow this best guess (~P2). Given how coarse our awareness of the possible consequences of a marginally higher economic growth rate in fact is, we indeed are in such a situation (~P3).
DiGiovanni also offers a comprehensive list of potential objections to his argument, and their rebuttals, but he mentions two attempts to find action-guidance despite cluelessness: i) consequentialist bracketing, and ii) a metaepistemic wager. Both are discussed below, right after we get more precise on DiGiovanni's empirical premise.
The Cross-Cause Fund (CCF) is an expert-managed cause-neutral fund. It is run by Rethink Priorities' cross-cause fund team, which is composed of members of its Leadership and Executive Team, Worldview Investigation Team, and Interdisciplinary Research Team.
It was launched in May 2026 as a fund explicitly aimed at full impartiality across(across all cause areas.areas)[1] — at a time when, per RP's own analysis, only 2% of EA prioritization resources were allocated to cross-cause prioritization, an area RP describes as historically underdeveloped in effective giving.
This sentence from the CCF website offers a good summary of the goal and intent:
With two caveats. First, RP cannot consider political candidates as potential beneficiaries of donations. Second, as of August 2026, multiple worthy causes have not yet been included in the evaluation, e.g., meta work (though the causes that have been included were prioritized intentionally, given specific cross-cause prioritization assumptions)assumptions plus lack of time to get closer to an ideal allocation). See the CCF FAQ for more related information.
Greaves, Hilary, and William MacAskill. 2025. “The Case for Strong Longtermism.” In Essays on Longtermism: Present Action for the Distant Future, edited by Hilary Greaves, Jacob Barrett, and David Thorstad. Oxford University Press. https://doi.org/10.1093/9780191979972.003.00030003..
Impartial Partially. 2026. "Bottom-Up Bracketing: A Person-Centered Response to Cluelessness." October 17. https://forum.effectivealtruism.org/posts/qpQSEy7n2WjxazFnf/bottom-up-bracketing-a-person-centered-response-to
To differentiate it from metanormative bracketing.
The Cross-Cause Fund (CCF) is an expert-managed cause-neutral fund. It is run by Rethink Priorities' cross-cause fund team, which is composed of members of its Leadership and Executive team,Team, Worldview Investigation Team, and Interdisciplinary Research Team.
It was launched in May 2026 as a fund explicitly aimed at full impartiality across all cause areas.[1] — at a time when, per RP's own analysis, only 2% of EA prioritization resources were allocated to cross-cause prioritization, an area RP describes as historically underdeveloped in effective giving.
This sentence from the CCF website offers a good summary of the goal and intent:
I've now brought back and lightly edited this paragraph that gives a helpful example. I think all the rest had low relevance, was misleading, or outright false.
Many objections/reactions to DiGiovanni's argument would be worth discussing, but I'm worried about transforming this tag page into a long paper. So a lot of work is done by: