Pronouns: he/him
Leave me anonymous feedback: https://docs.google.com/forms/d/e/1FAIpQLScB5R4UAnW_k6LiYnFWHHBncs4w1zsfpjgeRGGvNbm-266X4w/viewform
Contact me at: johnmichaelbridge[at]gmail[dot]com
Epistemic status: Uncertain and speculative. I try not to caveat my claims too much because it makes everything harder to read. If I’ve worded something too strongly, feel free to ask for clarification.
Thanks for your message Jackson. A few thoughts:
First thing is that, if intersectionality seems vague or poorly defined, then that’s likely a fault of my writing rather than the idea. To clarify—“intersectionality” is the idea that different individuals encounter overlapping types of disadvantage, and that these disadvantages combine in ways cannot be easily explained by looking at either kind of disadvantage in isolation. This means that finding solutions to issues at the intersection of several axes of disadvantage often requires explicitly considering how these considerations interact.
Its (potential) relevance to EA comes from the fact that a lot of EA cause areas deal with multiple axes of disadvantage in tandem; the use of ‘intersectionality’ could help to bring conceptual clarity to these discussions. To return to the animal welfare example, evelyciara has highlighted that there have been several recent posts about non-human animals being neglected on several axes simultaneously. Given the multiple disadvantages faced by future animals, intersectionality predicts that we will need to come up with novel solutions to protect them. Just trying to do (a) protect the long-term future, and (b) promote animal welfare is unlikely to achieve this goal. Guy Raveh highlights a similar example below in global health. I think the language of intersectionality is a neat way to explain what’s going on here, and why we might need to bring a fresh approach to these issues.
I don’t think animal welfare is the only cause area where intersectionality could bring conceptual clarity and improve our thinking. For example, engaging with how best to advance the welfare of digital people might benefit from an intersectional framing. It seems plausible that digital societies might end up with similar social ills—status games, inequality, ‘poverty’, etc—that we currently suffer. However, it’s unlikely that standard EA development strategy (read: health interventions) would be at all useful in dealing with these issues. Again, that’s because this is an intersectional issue, with multiple disadvantages (digital, poor) combining to create novel problems. If you agree with me that this seems obvious, then I think our disagreement has more to do with the use of the particular term ‘intersectionality’. This brings me to my next point.
Even if intersectionality comes with intellectual baggage, I don’t think we should shy away from using the term if it improves clarity. EAs already use terms that come with significant ideological baggage, because they’re useful and help to express important ideas. The term ‘nonhuman animals’ is a good example here—EAs use it to indicate that the moral distinction between the two is illusory. But this term (and much of the language around veganism) is morally charged, indicating a set of beliefs is perceived by many outside of EA as an indictment of meat-eaters. Alternatively, EAs on the forum often discuss political liberalism or cosmopolitanism, and many leading EAs explicitly identify as neoliberals. All three terms are highly politically charged, identifying a fuzzily defined set of policy stances that are controversial on both sides of the political spectrum. Nonetheless, in all of the cases I’ve just outlined, we use these ideas because they’re a helpful way of concisely explaining our ideas. I don’t think intersectionality is different in any unique way from the terms I’ve just described. I now think that this comment is right, inasmuch as it’s worth starting a new language game given the baggage that comes with the term.
I think this covers most of your comments, but please let me know if there’s anything I can clarify. I expect our crux of disagreement is on how useful it is to introduce a politically charged term like intersectionality into EA discourse, and I’m happy to engage more on that topic.
One of the reasons I no longer donate to EA Funds so often is that I think their funds lack a clearly stated theory of change.
For example, with the Global Health and Development fund, I’m confused why EAF hasn’t updated at all in favour of growth-promoting systemic change like liberal market reforms. It seems like there is strong evidence that economic growth is a key driver of welfare, but the fund hasn’t explained publicly why it prefers one-shot health interventions like bednets. It may well have good reasons for this, but there is absolutely no literature explaining the fund’s position.
The LTFF has a similar problem, insofar as it largely funds researchers doing obscure AI Safety work. Nowhere does the fund openly state: “we believe one of the most effective ways to promote long term human flourishing is to support high quality academic research in the field of AI Safety, both for the purposes of sustainable field-building and in order to increase our knowledge of how to make sure increasingly advanced AI systems are safe and beneficial to humanity.” Instead, donors are basically left to infer this theory of change from the grants themselves.
I don’t think we can expect to drastically increase the take-up of funds without this sort of transparency. I’m sure the fund managers have thought about this privately, and that they have justifications for not making their thoughts public, but asking people to pour thousands of pounds/dollars a year into a black box is a very, very big ask.
Thanks for this Rory, I’m excited to see what else you have to say on this topic.
One thing I think this post is missing is a more detailed response to the ‘ideal governance as weird’ criticism. You write that ‘weird ideal governance theories may well be ineffective’, but I would suggest that almost all fleshed-out theories of ideal AI governance will be inescapably weird, because most plausible post-trasformative AI worlds are deeply unfamiliar by nature.
A good intuition pump for this is to consider how weird modern Western society would seem to people from 1,000 years ago. We currently live in secular market-based democratic states run by a multiracial, multigender coalition of individuals whose primary form of communication is the instantaneous exchange of text via glowing, beeping machines. If you went back in time and tried to explain this world to an inhabitant of a mediaeval European theocratic monarchy, even to a member of the educated elite, they would be utterly baffled. How could society maintain order if the head of state was not blue-blooded and divinely ordained? How could peasants (particularly female ones) even learn to read and write, let alone effectively perform intellectual jobs? How could a society so dependent on usury avoid punishment by God in the form of floods, plagues or famines?
Even on the most conservative assumptions about AI capabilities, we can expect advanced AI to transform society at least as much as it has changed in the last 1,000 years. At a minimum, it promises to eliminate most productive employment, significantly extend our lifetimes, allow us to intricately surveil each and every member of society, and to drastically increase the material resources available to each person. A world with these four changes alone seems radically different and unfamiliar to our own, meaning any theory about its governance is going to seem weird. Throw in ideas like digital people and space colonisation and you’re jumping right off the weirdness deep end.
Of course, weirdness isn’t per se a reason not to go ahead with investigation into this topic, but I think the Wildeford post you cited is on the right track when it comes to weirdness points. AI Safety and Governance already struggles for respectability, so if you’re advocating for more EA resources to be dedicated to the area I think you need to give a more thorough justification for why it won’t discredit the field.
Red-team—“Are longtermism and virtue ethics actually compatible?”
A convincing red-team wouldn’t need a complex philosophical analysis, but rather a summary of divergences between the two theories and an exploration of five or six ‘case studies’ where consequentialist-type behaviour and thinking is clearly ‘unvirtuous’.
Explanation—Given just how large and valuable the long-term future could be, it seems plausible that longtermists should depart from standard heuristics around virtue. For instance, a longtermist working in biosecurity who cares for a sick relative might have good consequentialist reasons to abandon their caring obligations if a sufficiently promising position came up at an influential overseas lobbying group. I don’t think EAs have really accepted that there is a tension here; doing so seems important if we are to have open, honest conversations about what EA is, and what it should be.
The Cannonball Problem:
Doing longtermist AI policy work feels a little like aiming heavy artillery with a blindfold on. We can’t see our target, we’ve no idea how hard to push the barrel in any one direction, we don’t know how long the fuse is, we can’t stop the cannonball once it’s in motion, and we could do some serious damage if we get things wrong.
Hi Chris, I’ve responded to this somewhat in my response to Jackson above.
FWIW, I’m trying to avoid focusing on race questions here, because I think they’re pretty charged and racial equality isn’t an EA cause area in any case. Still, I think it’s worth responding to your comment that:
… the model doesn’t deal very well with the fact that Asians earn more on average than people who are white or an Asian woman may be better off in some ways than an Asian man (such as dating).
I actually think the model deals very well with this, as intersectionality would predict that being Asian + a minority + male would present a separate set of issues to being Asian + a minority + female. So fact that Asian women do better off in the dating scene is actually a pretty good example of intersectionality creating novel outcomes that might not be easily predicted by just stacking up disadvantages.
If, for some reason, discrimination against minority Asian women suddenly became an EA cause area, then people wanting to tackle the issue might do better trying tospecifically tackle ‘issues encountered by minority Asian women’ rather than just generally ‘reduce discrimination against women/minorities/Asians’.
Hi Tyner, thanks for your message,
I don’t have any studies I can point to on this, no, but the idea that privileged white men find it easier to take a universalising, impartial approach to doing good seems intuitively plausible. Admittedly, most of the data I have to support that argument are from private conversations, along with a general lack of demographic diversity in EA.
I’m open to the idea that I could be wrong here—can I ask you to explain in a little more detail why you feel that the PoC case isn’t unique?
~80% of the applications are speculative, from people outside the EA community and don’t even really understand what we do...
Out of interest—do you folks tend to hire outside the EA community? And how much does EA involved-ness affect your evaluation of applications?
I ask as I know some really smart and talented people working on development outside of EA who could be great founders, and I’d like to know if it’s worth encouraging them to apply.
I should clarify—by pocketing stuff, it ends up in the automatic queue for things to read. That way, I don’t really have to think about what to read next, and the things I want to read just pop up anyway.
Also strong upvote. I think nearly 100% of the leftist critiques of EA I’ve seen are pretty crappy, but I also think it’s relatively fertile ground.
For example, I suspect (with low confidence) that there is a community blindspot when it comes to the impact of racial dynamics on the tractability of different interventions, particularly in animal rights and global health.[1] I expect that this is driven by a combination of wanting to avoid controversy, a focus on easily quantifiable issues, the fact that few members of the community have a sociology or anthropology background, and (rightly) recognising that every issue can’t just be boiled down to racism.
Strong upvote—this is a really great post and helped me understand the source of many disagreements between myself and my more social justice-oriented friends.
I think you might have misunderstood the scope of this post. I want to emphasise that I endorse none of the following claims:
[a white person can never] understand or model the discrimination or pain faced by a (for example) queer, poor, Black and Muslim individual.
it’s pointless for a male to study female psychology because a male will never understand what it’s like to be female and should instead have no voice in the conversation
diversity and “equity” and “justice” [are] innate, disseminated values, rather than potential or circumstantial instrumental ones for prior lauded ends.
[EA] is somehow inherently in the wrong or discriminatory or evil for ending up mostly male, white, secular, tech based etc.
if your goal is to maximize diversity then you need intersectionality
you have to shame people for making assumptions or holding beliefs about those who are part of other [social] categories
If we remove these claims and just consider whether intersectionality would be a useful tool (of many different possible tools) for helping EAs think through difficult ideas, would this change your position at all?
That said, your comment has shifted me towards your perspective that intersectionality is unlikely to be useful for EAs, and it’s better to start a new language game. I think the word comes with enough baggage that it is hard to use as a neutral tool for analysing issues, and is liable to be misunderstood.
Thanks for helping to improve my thinking here.
Thanks for your message David. I think this probably depends on your definition of ‘discrimination’ - in SJ language, discrimination is typically something that happens at a systemic, rather than an individual level. That is to say, a set of policies that systematically disadvantage a particular group can still be discrimination if they reflect a prevailing system in which that group is disadvantaged. This can be true even if there is no bad intent on the part of individuals.
I think this broader definition is not always helpful, particularly because (a) it often fuels controversy to describe such policies as discriminatory, and (b) it is not always clear that any given policy is actually a part of a system of discrimination. That’s why I’ve tried to use ‘disadvantage’ in most of the post, which is a less loaded term. Nonetheless, I feel it’s the most appropriate term to use in the DeGraffenreid context, as the originators of Intersectionality Theory describe the case in terms of discrimination.
Is there a way you feel I could make this more clear?
I’m a bit late to the party on this one, but I’d be interested to find out how differential treatment of indigenous groups in countries where snakebites are most prevalent impacts the tractability of any interventions. I don’t have any strong opinions about how significant this issue is, but I would tentatively suggest that a basket of ‘ethnic inequality issues’ should be considered a third ‘prong’ in the analysis of why snakebites kill and maim so many people, and could substantially impact our cost-effectiveness estimates.
Explanation:
The WHO report linked by OP notes that, in many communities, over 3⁄4 of snakebite victims choose traditional medicine or spiritual healers instead of hospital treatment. I don’t think this is a result of either of the two big issues that the OP identifies—it doesn’t seem to stem from difficulty with diagnosis or cost of treatment, so much as being a thorny problem resulting from structural ethnic inequalities in developing countries.
I’m most familiar with the healthcare context of Amazonian nations, where deeply embedded beliefs around traditional medicine and general suspicion of mestizo-run governments can make it more difficult to administer healthcare to indigenous rainforest communities, low indigenous voter turnout reduces the incentives of elected officials to do anything about poor health outcomes, and discriminatory attitudes towards indigenous people can make health crises appear less salient to decisionmakers. Given that indigenous groups in developing countries almost universally receive worse healthcare treatment, and given that much indigenous land is in regions with high vulnerability to snake envenoming,[1] I wouldn’t be surprised if this issue generalised outside of Amazonia.
Depending on the size of the effect here, this could considerably impact assessments of tractability. For example, if developing country governments won’t pay for the interventions, it might be difficult to fund long-term antivenom distribution networks. Alternatively, if indigenous groups don’t trust radio communications, communicating health interventions could be particularly difficult. Also, given the fact that ‘indigenous’ is a poorly-defined term which refers to a host of totally unrelated peoples, it might be difficult to generalise or scale community interventions.
Study here (which I’ve not read).