Non-Western EAs’ perception of cross cultural interactions they had with Western EAs

By Yi-Yang @ 2024-07-24T02:59 (+72)

Summary

I investigated non-Western EAs’ perception of cross cultural interactions (CCIs) they had with Westerners, specifically:

  1. Whether or not non-Westerners experienced CCI issues, and how often;
  2. How their CCIs compare between EA and non-EA settings;
  3. What kinds of subtle acts of exclusion (SAEs) they had experienced.

I interviewed 21 non-Western EAs (selected from an EA conference’s Swapcard and a few from my own personal network) and discovered:

Methodology

I thought a more hands-on qualitative approach, like doing interviews, would be a better choice compared to a survey, because it offered me:

  1. More flexibility to pivot the type of questions I ask or the things I want to say;
  2. More information about a person’s emotional state;
  3. A way to potentially express empathy to those who might need it.

I’ve also received feedback that interviewing people seems like the next best option too.

Hence, I decided to interview people online who would identify themselves as EA or EA adjacent, and are predominantly non-Western. In these interviews, I asked:

  1. How much cross cultural interactions in EA have you had?
  2. How are the cross cultural interactions in EA settings that you’ve experienced?
  3. Have you encountered any kinds of subtle acts of exclusion from others in EA settings?
  4. Have you encountered acts of exclusion that are done by oppressed groups or minorities onto themselves in EA settings?
  5. How do your cross-cultural experiences compare between EA and non-EA settings?
  6. Are there other experiences you’d like to share? Or questions you'd like me to ask but I didn’t?

I did two things with the qualitative data I got from the interviews:

  1. I collected their experiences, paraphrased them, and compiled them under the appendix below. For those I found to be resonant in some hard-to-describe way, I included them in the “highlighted negative/positive vignettes” sections.
  2. I did some basic qualitative research coding and analysis to figure out how often similar experiences occur. You can find these in the “results” section.

To find the people I interviewed, I did the following:

  1. I selected 73 profiles from an EA conference’s Swapcard. I found them on meetups that non-Westerners generally attend, and on the attendees list (using the “country” filter).
  2. Then, I selected 43 out of 73 profiles to reach out to. These are the people who I had identified as having studied in, worked in, and lived in a non-Western country for the majority of their lives. I made some guesswork here by reading through their LinkedIn profiles.
  3. Of the 43 people, 21 of them did a 30-minute interview with me. Furthermore, I interviewed three additional people outside of the pool of EA conference attendees, for a total of 24 people.
  4. 3 interviewees did not review my paraphrases and did not give me consent to publish, so the final number of vignettes is 21.

Results

1.  Do non-Westerners experience CCI issues? And how many?

An overwhelming number of interviewees (19 out of 21) found their CCIs in EA settings to be almost all neutral or positive.

However, among the same 19 interviewees who found their CCIs to be mostly neutral or positive, they’ve also reported the following:

2. How do CCIs compare between EA and non-EA settings?

When asked to compared CCIs between EA settings and non-EA settings,

3. What kinds of SAEs do interviewees experience?

Highlighted negative vignettes

I thought it would be great to highlight some vignettes that felt resonant to me. I wanted to include more, but I felt like readers would appreciate a more condensed writing. If you have the time, you can read more of these vignettes in the Appendix.

Note: I used random English alphabets to anonymise the interviewees' names.

  1. In a workshop related to an EA related cause area, Western and non-Western folks sat at separate tables, probably unintentionally. Someone B knows who attended a similar workshop at a different time had a similar experience.
  2. In one of the afterparties during an EA conference, K had two potentially negative CCIs:
    1. When they went in the wrong direction, someone yelled the right direction to them in a way that felt infantilising and demeaning.
    2. When they were inside the afterparty space, no one seemed interested in engaging with them, so they left early.
  3. L had a call with an EA grantmaker. The purpose of the call was to receive feedback about a grant application they were rejected from. L thought that the delivery of the feedback was callous and the grantmaker's body language was giving off a "I don't want to be here" vibe. The already-supposed-to-be-short call also ended much earlier than expected. L felt uncertain about what factors contributed to that dynamic, but they suspected the difference in culture may play a part.
  4. Although there were a few Africans in an EA conference in Europe, L still felt a little suspicious that they didn’t see any African Europeans.

    Personal notes from Yi-Yang: I think the message L is trying to convey here is that it’s weird that they did not see any African Europeans in an EA conference in Europe, when they expect there to be more African Europeans than (Continental) Africans. I suspect L might be worried that African Europeans found EA circles in Europe to be unwelcoming.
     
  5. In conversations with Western EAs in EA conferences, M sometimes feels like they’re treated like a lower-status implementer or information source instead of an equal-status collaborator.

    Personal notes from Yi-Yang: Concretely, this might look like:
    - Assuming a person is ignorant and not asking whether they know something.
    - Giving off a body language that the person they’re speaking to is not competent enough.
     
  6. In a separate discussion, M noticed that their opinions were repeatedly criticised by another Western EA, which prompted them to stop engaging in the discussion. This felt abnormal even by EA's higher truth-seeking norms and this made them feel suspicious that there are other factors at play here (namely their ethnicity/culture and gender).
  7. R noticed at least one discussion around how remuneration should be calculated between different countries had overtones of “the cost of living is too high in higher income countries”, which failed to “capture trade-offs around infrastructure, pollution, and social trust” in low-to-middle-income countries. They think discussions around fair remuneration for talent should not only consider the cost of living in a city and purchasing power parity, but also other factors too.
  8. S observed that a Western EA expressed their intention to help some non-Western non-EA locals with an issue but assumed that the locals lacked the appropriate concern or knowledge to tackle said issue. S felt (low-to-medium confidence) that this person was quite patronising and arrogant towards the locals.

    Personal notes from Yi-Yang: Imagine a highly educated person from a high income country going to a low-to-middle-income country. They noticed some potential problems and wanted to help the locals, but assumed they were ignorant about the problem.
     
  9. Despite not being a native English speaker themself, T would sometimes unconsciously desire less interactions with folks with poor English language skills. T suspects that they remind them of themselves when they had a poor grasp on English, which they associate with their feelings of inferiority at the time, and which they project onto the subjects.
  10. V attended an AI safety event where the organiser mentioned how the global south isn’t that important for AI safety. V thought that what the organiser said could be a reasonable stance, but the way it was phrased felt quite icky to them. It also didn’t help that the organiser’s attempt at sugarcoating their stance made it more icky.

    Personal notes from Yi-Yang: I’m not sure how they were exactly sugarcoating their stance, but it might look like: “there are other EA cause areas that are more important in the global south.”

Highlighted positive vignettes

  1. K had two positive CCIs during an EA conference:
    1. A Western EA invited them unprompted for a drink.
    2. In a group house, they found themselves being included in conversations more often.
  2. R thinks that the CCIs they had experienced were mostly positive.
    1. In fact, they wished EAs were less worried about offending them.
    2. They also had helpful meta-conversations to deconfuse certain non-verbal cues.
    3. R also liked that there’s much less pressure to drink alcohol in EA settings.
  3. W thought the socioeconomic diversity meetup [in an EA conference] got them to meet similar folks and made them feel more welcomed.
  4. X remembered a positive experience where two Westerners recognised that they were dominating a discussion and then decided to speak less so that X could be included.
  5. Despite cultural barriers, Z thought that their CCIs are generally respectful. They especially like how diverse their team is in an EA aligned organisation.

If you have the time, you can read more of these vignettes in the Appendix.

Highlighted neutral vignettes

  1. R has a bigger preference for bonding with people who are culturally similar to them. The side-effect of excluding those who are culturally dissimilar seems like the right trade-off at times, because hanging out with people who are similar is more comforting and freeing.

CCIs interviewees reported and their frequency

A. General negative CCIs

CCIs interviewees reportedNumber of interviewees who reported this
Non-Western EAs found the act of connecting with Western EAs challenging.4
Non-Western EAs felt suspicious about the lack of representation.3
Non-Western EAs found the English language barrier challenging to overcome.3
Non-Westerner EAs found the application process for EA-related programmes based in Western countries to be daunting.2
Western EAs did not weigh the cheaper costs of living in low-to-middle-income countries (LMICs) with the less tangible costs of living in LMICs (e.g., poor infrastructure, pollution, etc) when thinking about remuneration.1
Western EAs paid less attention to or ignored non-Western EAs due to language barriers or accent.1
Western EAs were persuading folks to adopt non-EA-but-common-among-EAs norms (like polyamory) in EA spaces a little too excessively.1
Non-Western EAs did not engage in EA conferences in Western settings because they felt uncomfortable and overwhelmed.1
Non-Western EAs felt stressed about sounding smart and competent when talking to Western EAs.1

B. SAEs caused by Western EAs

CCIs interviewees reportedNumber of interviewees who reported this
Western EAs treating non-Western EAs in a way that's demeaning.4
Western EAs were coming across as paternalistic towards non-Western EAs.2
Western EAs mentioned that the global south isn't important for AI safety in a way that's demeaning.1
Western EAs did not include non-Western EAs in discussions or conversations when in predominantly Western settings.1
Western AI Safety (AIS) researchers believed that the primary intention of Chinese AIS researchers is to gain status and not to improve AIS.1
Western EAs treat non-Western EAs as if they're lower status.1
Western EAs initially changed their behaviour to accommodate non-Western EAs, but after a while, reverted to prior behaviour that's less accommodating.1
Discussions with Western EAs about the intersection of their country and AI were a little too antagonistic or epistemically poor.1
English is the primary language used in their local group when they prefer their own native language.1
A non-Western cuisine in an EA event was introduced with the wrong country of origin.1

C. SAEs caused by non-Western EAs

CCIs interviewees reportedNumber of interviewees who reported this
Non-Western EAs changing their accent or communication style to be more Western.2
Non-Western EAs unconsciously desired less interactions with non-Western EAs with poor English language skills.1
Non-Western EAs avoided food from where they're from.1
Non-Western EAs pushed back against Western EAs less, but pushed back against non-Western EAs more.1

Demographics of interviewees

Gender

I did not ask my interviewees what gender they identify with. However, just from my engagement with them, I suspect there are 10 female-identifying interviewees and 11 male-identifying interviewees.

Locale

I did not ask my interviewees where they are from and where they are living now. I mostly inferred this information from their LinkedIn profiles and from a bit of small talk during the interview.

Geographical subregion according to UN geoscheme / ContinentMostly lived inNow living in
Eastern Asia53
South-eastern Asia95
Southern Asia41
Africa*21
South America11
Northern America01
Northern Europe08

* It might be easy to identify the interviewees if I specified the African subregions they are from, so I chose to put them all in one continent.

Limitations

  1. There are too many variables to control for (e.g., social status differences, personality, social skills, my own interview skills, etc), and I expect the qualitative data I collected to be noisy.
  2. There’s probably some selection bias--most of my interviewees are also attendees from an EA conference, and I wasn’t being extremely systematic with who I selected.
  3. The “Western” and “non-Western” categorisation can be somewhat arbitrary--there may be folks who are somewhere in the middle that were not selected for.
  4. Without even considering how memory retrieval can sometimes be a hit-or-miss, SAEs are already hard to detect.
    1. I expect there to be misinterpretations of what actually happened by my interviewees (e.g., some people could just be rude to everyone and it had nothing to do with differences in culture). And I don’t think I did a good job of steering interviewees towards giving me empirical observation of what happened.
    2. Giving more examples of SAEs during interviews would stretch the time limit. And I noticed my interviewee’s scope of what SAEs can be were limited to the examples I gave. So I expect some under-reporting going on.
  5. I sometimes nudged my interviewees to think about other possible interpretations. Probably because I have a (possibly mistaken) worry that readers might think that my interviewees are not thoughtful enough and hence their vignettes should be dismissed as unreliable. This might have caused my interviewees to think that only experiences above a certain threshold should be shared, and hence made them less likely to share more.
  6. I was too focused on negative CCIs and forgot to include questions around positive CCIs.
  7. Some of the vignettes can come across as lacking concrete details, but I did that to protect my interviewees’ anonymity.
  8. I’m a Chinese from South-eastern Asia and I noticed a bias towards interviewing more folks from Southeast Asia and East Asia, which isn’t very representative of actual demographics in the EA community.

Appendix

Definition of subtle acts of exclusion

Other notes:

  1. SAE can be done by someone else unto the victim, or done by the victim unto the victim themself.
  2. SAE can happen either intentionally or unintentionally; and consciously or unconsciously
  3. The term “subtle acts of exclusion” came from Jana and Baran’s book of the same title, and I’ve made one major modification to the term.

Vignettes

B

C

D

F

G

H

J

K

L

M

N

P

Q

R

S

T

V

W

X

Y

Personal notes from Yi-Yang: I suspect a combination of language barriers and cultural barriers made the usual overwhelmingness of a conference, even more overwhelming.

Z

Joseph Lemien @ 2024-07-24T23:04 (+4)

It can be tricky to explore some of these topics that overlap between cultural background, nationality, how others perceive us, differing norms, assumptions, and communication styles.  It can be hard to parse between what we view as the gradient between reasonable and unreasonable assumptions.[1]

Nonetheless, I also really like that there are people in this community who notice and who are aware of subtle things. 

I'm glad to read explorations of these kinds of things, and I'm glad that you've spent all this time and effort exploring it and sharing some of your findings. Thank you.

  1. ^

    Such as tentatively predicting that a Black American in Chengdu probably doesn't grow up speaking Mandarin, as opposed to very confidently assuming that an Asian of unknown nationality in Chicago couldn't possibly have grown up speaking English

Yi-Yang @ 2024-07-25T02:24 (+3)

It can be hard to parse between what we view as the gradient between reasonable and unreasonable assumptions

Thanks for bringing this up, I think I did significantly update to being a bit more okay with folks being ignorant, or having false assumptions. It's tough! I would probably make the same mistakes too and would want some space to fumble around and correct myself.

Joseph Lemien @ 2024-07-25T12:23 (+5)

space to fumble around

Having that space/slack that strikes me as very valuable. Maybe it is somewhat analogous to the idea of psychological safety that gets talked about so often when discussing company culture.

Some of my most frustrating inter-personal situations involved me being judged harshly for getting something wrong in which (in my perspective) it was totally reasonable and understandable for me to make a guess and get it wrong.