It's worth noting that the 2017 EA Survey (data collected but not yet published), the 2015 EA Survey, and the 2014 EA Survey all have global poverty as the most popular cause (by plurality) among EAs in these samples, and by a good sized margin. So it may not be the case that the EA movement is misrepresenting itself as a movement by focusing on global poverty (even if movement leaders may think differently than the movement as a whole).

Still, it is indeed the case that causes other than global poverty are more popular than global poverty, which would potentially argue for a more diverse presentation on what EA is about as you suggest.

Hi Michael,

I agree the issue of people presenting EA as about global poverty when they actually support other causes is a big problem.

80k stopped doing this in 2014 (not a couple of months ago like you mention), with this post: https://80000hours.org/2014/01/which-cause-is-most-effective-300/ The page you link to listed other causes at least as early as 2015: https://web.archive.org/web/20150911083217/https://80000hours.org/articles/cause-selection/

My understanding is that the GWWC website is in the process of being updated, and the recommendations on where to give are now via the EA Funds, which include 4 cause areas.

These issues take a long-time to fix though. First, it takes a long time to rewrite all your materials. Second, it takes people at least several years to catch up with your views. So, we're going to be stuck with this problem for a while.

In terms of how 80,000 Hours handles it:

Their cause selection choices, which I think they updated a few months ago only really make sense if you adopt total utilitarianism (maximise happiness throughout history of the universe) rather than if you prefer a person-affecting view in population ethics (make people happy, don’t worry about creating happy people) or you just want to focus on the near future (maybe due to uncertainty about what we can do or pure time discounting).

This is a huge topic, but I disagree. Here are some quick reasons.

First, you should value the far future even if you only put some credence on theories like total utilitarianism.

e.g. Someone who had 50% credence in the person affecting view and 50% credence in total utilitarianism, should still place significant value on the far future.

This is a better approximation of our approach - we're not confident in total utilitarianism, but some weight on it due to moral uncertainty.

Second, even if you don't put any value on the far future, it wouldn't completely change our list.

First, the causes are assessed on scale, neglectedness and solvability. Only scale is affected by these value judgements.

Second, scale is (to simplify) assessed on three factors: GDP, QALYs and % xrisk reduction, as here: https://80000hours.org/articles/problem-framework/#how-to-assess-it

Even if you ignore the xrisk reduction column (which I think would be unreasonable due to moral uncertainty), you often find the rankings don't change that much.

E.g. Pandemic risk gets a scale score of 15 because it might pose at xrisk, but if you ignored that, I think the expected annual death toll from pandemics could easily be 1 million per year right now, so it would still get a score of 12. If you think engineered pandemics are likely, you could argue for a higher figure. So, this would move pandemics from being a little more promising than regular global health, to about the same, but it wouldn't dramatically shift the rankings.

I think AI could be similar. It seems like there's a 10%+ chance that AI is developed within the lifetimes of the present generation. Conditional on that, if there's a 10% chance of a disaster, then the expected death toll is 75 million, or 1-2 million per year, which would also give it a score of 12 rather than 15. But it would remain one of the top ranked causes.

I think the choice of promoting EA and global priorities research are even more robust to different value judgements.

We actively point out that the list depends on value judgements, and we provide this quiz to highlight some of the main ones: https://80000hours.org/problem-quiz/

Effective Altruism is quite difficult to explain if you want capture all of its complexity. I think that it is a completely valid choice for an introductory talk to focus on one aspect of Effective Altruism as otherwise many people will have trouble following.

I would suggest letting people know that you are only covering one aspect of Effective Altruism, ie. "Effective Altruism is about doing the most good that you can with the resources available to you. This talk will cover how Effective Altruism has been applied to charity, but it is worth noting that Effective Altruism has also been applied to other issues like animal welfare or ensuring the long-term survival of humanity".

This reduces the confusion when they hear about these issues later and reduces the chance that they will feel mislead. At the same time, it avoids throwing too many new ideas at a person at once which may reduce their comprehension and it explains how it applies to an issue which they may already care about.

I broadly agree with you on the importance of inclusivity, but I’m not convinced by your way of cashing it out or the implications you draw from it.

Inclusivity/exclusivity strikes me as importantly being a spectrum, rather than a binary choice. I doubt when you said EA should be about ‘making things better or worse for humans and animals but being neutral on what makes things better or worse’, you meant the extreme end of the inclusivity scale. One thing I assume we wouldn’t want EA to include, for example, is the view that human wellbeing is increased by coming only into contact with people of the same race as yourself.

More plausibly, the reasons you outline in favour of inclusivity point towards a view such as ‘EA is about making things better or worse for sentient beings but being neutral between reasonable theories of what makes things better or worse’. Of course, that brings up the question of what it takes to count as a reasonable theory. One thing it could mean is that some substantial number of people hold / have held it. Presumably we would want to circumscribe which people are included here: not all moral theories which have at any time in the past by a large group of people are reasonable. At the other end of the spectrum, you could include only views currently held by many people who have made it their life’s work to determine the correct moral theory. My guess is that in fact we should take into account which views are and aren’t held by both the general public and by philosophers.

I think given this more plausible cashing out of inclusivity, we might want to be both more and less inclusive than you suggest. Here are a few specific ways it might cash out:

• We should be thinking about and discussing theories which put constraints on actions you’re allowed to take to increase welfare. Most people think there are some limits on be what we’re allowed to do to others to benefit others. Most philosophers believe there are some deontological principles / agent centred constraints or prerogatives.

• We should be considering how prioritarian to be. Many people think we should give priority to those who are worst off, even if we can benefit them less than we could others. Many philosophers think that there’s (some degree of) diminishing moral value to welfare.

• Perhaps we ought to be inclusive of views to the effect that (at least some) non-human sentient beings have little or no moral value. Many people’s actions imply they believe that a large number of animals have little or no moral value, and that robots never could have moral value. Fewer philosophers seem to hold this view.

• I’m less convinced about being inclusive towards views which place no value on the future. It seems widely accepted that climate change is very bad, despite the fact that most of the harms will accrue to those in the future. It’s controversial what the discount rate should be, but not that the pure time discount rate should be small. Very few philosophers defend purely person-affecting views.

Interesting read!

Just a thought: does anyone have any thoughts on religion and EA? I don't mean it in a "saving souls is cost effective" way, more in the moral philosophy way.

My personal take is that unless someone is really hardcore/radical/orthodox, then most of what EA says would be positive/ethical for most religious persons. That is certainly my experience talking to religious folks, no one has ever gotten mad at me unless i get too consequentalist. Religious people might even be more open to giving what we can pledge, and EA altruism in some ways, because of the common practise of tithing. though they might decide that "my faith is the most cost effective", but that only sometimes happens, they seem to donate on top of it usually.

PS: Michael, was it my question on the facebook mental health EA post that prompted you to write this? Just curious.

Thank you for the interesting post, and you provide some strong arguments for moral inclusivity.

I'm less confident that the marketing gap, if it exists, is a problem, but there may be ways to sell the more 'weird' cause areas, as you suggest. However, even when they are mentioned, people may still get the impression that EA is mostly about poverty. The other causes would have to be explained in the same depth as poverty (looking at specific charities in these cause areas as well as cost-effectiveness estimates where they exist, for instance) for the impression to fade, it seems to me.

While I do agree that it's likely that a marketing gap is perceived by a good number of newcomers (based solely on my intuition), do we have any solid evidence that such a marketing gap is perceived by newcomers in particular?

Or is it mainly perceived by more 'experienced' EAs (many of whom may prioritise causes other than global poverty) who feel as if sufficient weight isn't being given to other causes, or who feel guilty for giving a misleading impression relative to their own impressions (which are formed from being around others who think like them)? If the latter, then the marketing gap may be less problematic, and will be less likely to blow up in our faces.

Hey Michael, sorry I am slightly late with my comment.

To start I broadly agree that we should not be misleading about EA in conversation, however my impression is that this is not a large problem (although we might have very different samples).

I am unsure where I stand on moral inclusivity/exclusivity, although as I discuss later I think this is not actually a particularly major problem, as most people do not have a set moral theory.

I am wondering what your ideal inclusive effective altruism outreach looks like?

I am finding it hard to build up a cohesive picture from your post and comments, and I think some of your different points don't quite gel together in my head (or at least not in an inconvenient possible world).

You give an example of this of beginning a conversation with global poverty before transitioning to explaining the diversity of EA views by:

Point out people understand this in different ways because of their philosophical beliefs about what matters: some focus on helping humans alive today, others on animals, others on trying to make sure humanity doesn't accidentally wipe itself out, etc.

For those worried about how to ‘sell’ AI in particular, I recently heard Peter Singer give a talk when he said something like (can't remember exactly): "some people are very worried about about the risks from artificial intelligence. As Nick Bostrom, a philosopher at the University of Oxford pointed out to me, it's probably not a very good idea, from an evolutionary point of view, to build something smarter than ourselves." At which point the audience chuckled. I thought it was a nice, very disarming way to make the point.

However trying to make this match the style of an event a student group could actually run, it seems like the closet match (other than a straightforward into to EA event) would be a talk on effective global poverty charity, follow by an addendum on EA being more broad at the end. (I think this due to a variety of practical concerns, such a there being far more good speakers and big names in global poverty, and it providing many concrete examples of how to apply EA concepts etc.)

I am however skeptical that a addendum on the end of a way would create nearly as strong an impression as the subject matter of the talk itself, and people would still leave with a much stronger impression of EA as being about global poverty than e.g. x-risk.

You might say a more diverse approach would be to have talks etc. roughly in proportion to what EAs actually believe is important, so if to make things simple, a third of EAs thought Global poverty was most important, a third x-risk and a third animal suffering, then a third of the talks should be a global poverty, a third on x-risk etc. Each of these could then end with this explanation of EA being more broad etc.

However if people's current perception that global poverty events is best way to get new people into EA is in fact right (at least in the short term) either by having better attendance or conversion ratios this approach could still lead to the majority of new EAs first introduction to EA being through a global poverty talk.

This due to the previous problem of the addendum not really changing peoples impressions enough we could still end up with the situation you say we should want to avoid where:

People should not feel surprised about what EAs value when they get more involved in the movement.

I am approaching this all more from the student group perspective, and so don't have strong views on the website stuff, although I will note that my impression was that 80k does a good job on being inclusive, and GWWC is more of an issue with a lack of updates than anything.

One thing you don't particularly seem to be considering is that almost all people don't actually have strongly formed moral views that conform to one of the common families (utilitarian, virtue ethics etc.) so I doubt (but could be wrong, as there would probably be a lot of survivor bias in this) that a high percentage of newcomers to EA feel excluded by the current implicit assumptions that might often be made of e.g. future people matter.

I favour the idea of inclusivity, and being upfront about the different areas that are prioritised. I think within these areas, there are certain ideas that could be held back, that people might think are fairly unusual issues of consideration. However, this does also mean that if an idea is put forward it doesn't then become a priority in itself. For example, where factory farming is put forward, not factory farming isn't necessarily the most effective way to reduce harm, instead harm can be shifted onto non-factory farming, which is another form of animal exploitation (often with a different set of harms) that would then need to be addressed. However, at that point it might fall off the EA radar, or be a reduced cause option because of a perceived decline in suffering has meant another general cause is more important.

However, the impact of animal consumption on the environment and human health are also strong reasons for prioritising the animal issue, and are even exacerbated by non-factory farming, particularly on the issue of the environment. So it might appear that reduction of animal consumption overall ought to be the priority. Then it might be more realistic to say that EA is interested in reducing harm to other animals, but this can also frame the issue in the harms that take place, and this can be notable in terms of further normalising those harms through trading one off for another, particularly when some people in the animal movement are ideologically opposed to the responsible systems.

However, I think it is ok to introduce the idea of factory farming, as long as the variety of approaches within the animal movement that address this issue are articulated later on (within EAA). Other issues such as wild animal suffering are also given time here. However, this brings up the issue of inclusivity of ideas, and i think it is fairly recognised that EAA is dominated by utilitarian thinking on the issue of animal exploitation. I think this is a cause of some concern because it tends to lead to a certain interpretation of ideas that are consequently prioritised over others, seemingly because they might fit with the idea the group has of itself, rather than because they have been thoroughly evaluated.

It seems to begin with EAA drew its expertise from a not particularly diverse form of 'mainstream' thinking that exists within the animal movement, and henceforth struggled (or uncharitably has been disinclined) to include different perspectives. This could largely be the result of traditional leaders in the animal movement having little incentive to include areas they know little about, or that would suggest their work may need to be adjusted in order to be more amenable to a more inclusive EAA. This also includes the concern that other perspectives may become increasingly favoured and their position diminished, or their favoured groups receive less funding overall.

In this sense an ingroup / outgroup situation can be created and perpetuated. So i would say the idea of being more inclusive is a good thing, but whether the idea of including is one that people want to engage in is another. In this sense it could highlight a previous (and ongoing) less than optimal approach, and this is quite a difficult situation to deal with. Both for people within EAA and for those who are interested in EAA but don't see their ideas included in conterfactual considerations, instead they might be generally ignored or dismissed because they don't fit with the view the organisation has constructed.

I think that going forward there need to be some changes in the way EAA works in order for it to grow, and for it to claim that it is indeed doing the most effective thing. As it stands there are too many ideologically similar groups whose ideas are prioritised and perpetuated and that have either resisted scrutiny, or the consequences of scrutiny. So it seems to me there is quite a challenge ahead if inclusivity is going to be the path people wish to take.

I totally understand your concern that the EA movement is misrepresenting itself by not promoting issues proportional to their representation among people in the group. However, I think that the primary consideration in promoting EA should be what will hook people. Very few people in the world care about AI as a social issue, but extreme poverty and injustice are very popular causes that can attract people. I don't actually think it should matter for outreach what the most popular causes are among community members. Outreach should be based on what is likely to attract the masses to practice EA (without watering it down by promoting low impact causes, of course). Also, I believe it's possible to be too inclusive of moral theories. Dangerous theories that incite terrorism like Islamic or negative utilitarian extremism should be condemned.

Also, I'm not sure to what extent people in the community even represent people who practice EA. Those are two very different things. You can practice EA, for example by donating a chunk of your income to Oxfam every year, but not have anything to do with others who identify with EA, and you can be a regular at EA meetups and discussing related topics often (ie. a member of the EA community) without donating or doing anything high impact. Perhaps the most popular issues acted on by those who practice EA are different from those discussed by those who like to talk about EA. Being part of the EA community doesn't give one any moral authority in itself.

This post may add grist to the mill that any such gap is a problem: https://srconstantin.wordpress.com/2017/01/11/ea-has-a-lying-problem/

(The post doesn't quite cover the same issues that Michael talks about here, but there's a parallel)