I’m currently a co-director at EA Netherlands (with Marieke de Visscher). We’re working to build and strengthen the EA community here.
Before this, I worked as a consultant on urban socioeconomic development projects and programmes funded by the EU. Before that, I studied liberal arts (in the UK) and then philosophy (in the Netherlands).
Hit me up if you wanna find out about the Dutch EA community! :)
I imagine you’d organise it the same way you’d organise any other national democratic organisation in the US—through representative structures, regional chapters, online participation options, and other standard approaches that democratic organisations use to manage scale and geography.
I asked Claude for examples:
The American Medical Association has around 270,000 members across all states and manages democratic governance through state medical societies that send delegates to their annual House of Delegates meeting, plus online voting for leadership positions. Professional engineering societies like IEEE operate similarly with over 400,000 members globally—they use regional sections, online balloting for board elections, and hybrid conferences. Even academic organisations like the American Psychological Association coordinate democratic decision-making across their 120,000+ members through divisional representation and electronic voting systems.
Thank you!
I’m sure members typically have a suboptimal level of organisational context. However, that might be outweighed by other benefits. I think the more important question is, ‘Are there contexts in which democratic decision-making processes improve outcomes?’. I would love it if people could point me to some good research on the subject!
Edit: after asking Claude to do some research, the best I could find was this.
There is much enthusiasm among scholars and public administrators for participatory and collaborative modes of governance as a means to tackle contemporary environmental problems. Participatory and collaborative approaches are expected to both enhance the environmental standard of the outputs of decision-making processes and improve the implementation of these outputs. In this article, we draw on a database of 305 coded published cases of public environmental decision-making to identify key pathways via which participation fosters effective environmental governance. We develop a conceptual model of the hypothesized relationship between participation, environmental outputs, and implementation, mediated by intermediate (social) outcomes such as social learning or trust building. Testing these assumptions through structural equation modeling and exploratory factor analysis, we find a generally positive effect of participation on the environmental standard of governance outputs, in particular where communication intensity is high and where participants are delegated decision-making power. Moreover, we identify two latent variables—convergence of stakeholder perspectives and stakeholder capacity building—to mediate this relationship. Our findings point to a need for treating complex and multifaceted phenomena such as participation in a nuanced manner, and to pay attention to how particular mechanisms work to foster a range of social outcomes and to secure more environmentally effective outputs and their implementation.
I’m somewhat confused about what led you to this conclusion. I was the co-director of EA Germany for two years, an organization that is similarly structured. When I compare it to the memberless nonprofits where I’m a board member, the overhead for organizing a general assembly has been greater, yet it hasn’t resulted in significant decision-making input from the members.
The post’s main claim is relatively modest: “you don’t need to panic about democracy in EA.” Speaking for myself, I contributed to this post because I have the impression that often, when someone suggests increased democratisation, the responses are mostly, “oh, that will never work because of this, that, and the other reason” before moving on. In writing this post, I wanted to update people away from that by providing an example of where democratic elements have worked reasonably well. Nonetheless, I agree that there will be examples of organisations where it doesn’t work as well, and maybe EA Germany is one of those organisations.
In the rest of your comment, it feels like you’re mainly questioning the value of national community building orgs rather than the value of democratising national community building orgs. That’s a reasonable thing to question, but I think it’s a separate discussion. Unless I’m misunderstanding you?
I agree!
This is cool—thanks for sharing!
Maybe you’ve already seen this Tom but, if not, the Rethink Priorities team published some findings from testing different framings of EA and longtermism here. The 2024 pulse also has a few interesting reports that may be useful. But I agree, if there’s more research it’d be super useful to see it.
Thanks, Emma! And sorry for the rather curt comment. I typed it out while my phone was dying and I was on the move after a rough start to the day. In hindsight, it came across more strongly than I intended.
I really appreciate the context you shared. To give a bit of background from my side: it’s felt like things are moving quite quickly on CEA’s end with branding and related work (which is exciting!), but from where I sit, it’s sometimes felt like orgs like ours are struggling to keep up. I can’t speak for all community builders, but I usually try to follow CEA’s lead to keep things coherent across the ecosystem — so the lack of engagement has at times felt a little disorienting, and honestly, a bit disempowering.
For example, we’ve recently been putting quite a bit of time and money into revamping our website and visual identity. If we’d known more about what was happening at CEA, it could have helped shape our direction. A short note to CEA’s groups team, which could have been forwarded on, would have gone a long way — something like: “Hey, we’re working on EA brand stuff at the moment. It’s being informed by such-and-such findings. Here’s our rough timeline and broad direction. More soon”.
What’s made that feeling a bit more acute is that I’d actually tried to reach out a couple of times to find someone at CEA to talk with about this, but didn’t get a response. I almost certainly didn’t go about it the right way (in hindsight, I probably should’ve just emailed you directly), but that lack of feedback added to the sense of being out of the loop.
That said, I’m really glad to hear about the progress and new capacity on your team, and I’d love to stay in sync however we can. And sorry once again for the abrasive tone of my comment!
I’ve been thinking about creating an EA Reddit account to contribute to EA-related discussions on there. This has prompted me to get round to it finally.
Great stuff! And thanks for taking the time to share what you’ve done.
Do you know if the team working on the EA brand project would be up for talking with professional community builders? At EA Netherlands we’re working on our brand quite a bit at the moment, and I think a few other national organisations are too. Since national orgs are often the main entry point for EA in their region, I think this should probably be done in coordination with CEA to make sure we’re all aligned.
To speak frankly, I’m a little surprised professional community builders haven’t been involved in brand work so far. (This comment isn’t addressed to you Agnes, from what I understand it’s not your responsibility to keep brand stakeholders in the loop! Writing it here in case a relevant person reads it).
Again, thanks for this work, it looks great!
Thanks!
Very cool! Thanks for sharing.
I’m curious, what proportion of your readers does Mailchimp report as being based in Amsterdam? I’m sure it’s quite misleading for various technical reasons but I’m still intrigued.
I heard from the EA Survey team that Amsterdam was the sixth largest hub, in terms of respondents, in the 2024 survey (2.4%). The biggest hub was London, which generated something like 6% of respondents.
You’re welcome :)
I suspect it’s mostly the way you’ve written it. As a rule of thumb, always aim for high reasoning transparency. I asked ChatGPT o3 to rewrite it in a style that’s more likely to appeal to EAs and that frames it in terms of reducing the risk of stable totalitarianism. I’ve pasted its output below.
”Claim. A second Trump presidency would raise the probability that the United States drifts toward a technologically-entrenched autocracy, thereby increasing the global risk of stable totalitarianism — a scenario where an oppressive regime locks in power for centuries or more.¹
Why that matters. Even a <1 % chance of permanent totalitarian lock-in constitutes an existential risk: it would foreclose almost all future value while inflicting vast suffering.² Emerging tech — especially frontier AI, ubiquitous surveillance, and autonomous weapons — could remove the usual checks (elite defection, popular uprising, leadership succession) that historically topple dictatorships.³
Mechanisms by which Trump plausibly raises the risk:
• Erosion of democratic guard-rails. Intent to purge the civil service and use federal agencies for partisan aims weakens the institutions that normally resist autocratic consolidation.
• Politicised AI and surveillance. Allies have floated centralising control of federal datasets and AI models; misused, these tools could neutralise opposition and entrench rule.
• Norms against power transfer. Open refusal to accept electoral defeat in 2020 signals willingness to test the limits of constitutional constraint.
Scale & neglectedness. The U.S. controls ~25 % of world GDP and a decisive share of AI R&D; trajectory changes here propagate globally. Yet only ~$70 m/yr flows to non-partisan democracy-protection charities, versus >$10 bn in partisan spend.
Tractability. Cost-effective levers include:
State-level democracy infrastructure (voter-registration, local media fact-checks) — historical cost ≈ $300–400 per net vote.
Legal defence funds for civil-service whistle-blowers.
AI-governance policy work that limits executive control over surveillance and autonomous-weapon deployment.
Next steps for funders / organisers: commission a quick Rethink Priorities dive to refine the risk delta; pilot $1–5 m to the most effective democracy-protection orgs; reassess post-election.
¹ See 80,000 Hours problem profile on risks of stable totalitarianism. 80,000 Hours
² 80k’s BOTEC puts the century-level risk at ~0.3 %, with other experts’ estimates up to 5 %. 80,000 Hours
³ Advanced AI could give a ruler decisive military, surveillance and succession advantages, removing historic failure modes for dictatorships.”
You’re right to flag the risks of introducing pay gates. I agree it would be a mistake to charge for things that are currently core to how people first engage, especially given how many people first get involved in their 20s when finances are tight.
I think the case for a supporter membership model rests on keeping those core engagement paths free (intro courses, certain events, 1-1 advice, etc.), while offering membership as an optional way for people to express support, get modest perks, and help fund infrastructure.
I also think the contrast you draw between the two (mountaineering clubs = self-benefit, EA = other-benefit) is too simplistic. Most people who get involved in EA do so because they want to become more effective at helping others. That’s a deeply personal goal. They benefit from gaining clarity, support, and a community aligned with their values. EA resources serve them, not just the ultimate beneficiaries.
Likewise, mountaineering clubs aren’t purely self-serving either — they invest in safety standards, trail access, training, and other mountaineering public goods that benefit non-members and future members.
In both cases, people pay to be part of something they value, which helps them grow and contribute more, and then the thing they value ends up growing as well.
I think you might be overestimating how much the NKBV offers as part of the basic membership. Most of their trips and courses, etc., are paid add-ons. What the €50 fee actually gets you is fairly lightweight: a magazine, eligibility to join trips (not free), discounted access to mountain huts (because the NKBV helps fund them), inclusion in their group insurance policy, and a 10% discount with a Dutch outdoor brand.
That’s not nothing, but it’s modest and it shows that people will pay for affiliation, identity, and access to a community infrastructure, even if the tangible perks are limited.
The EA equivalent could be things like discounted or early access to EAG(x) events, member-only discussion groups, or eligibility to complete advanced courses offered by national EA associations. If multiple countries coordinated, pooled membership fees could help subsidise international EA public goods such as the Forum, EAG(x) events, group support infrastructure, etc.
I think the key point is this: the NKBV shows that people are willing to pay for affiliation, even if the direct perks are modest, as long as the organisation feels valuable to their identity and goals. EA can plausibly do the same.
You’re right that EA’s current meta budget works out to far more than €50 per “involved person”—but that average includes the highly engaged core: people attending conferences, receiving 1-1 support, travel grants, and significant staff time.
A low-touch “supporter” tier is a different product entirely. If you ask someone for €50/year just to back the mission, receive a newsletter, and gain access to certain events, the marginal cost is minimal: card fees, a CRM entry, an occasional email, maybe a €5 welcome sticker. Even doubling every line item puts the cost at €10–20, leaving €30–40 net per person.
We could keep the high-cost, high-impact activities funded by major donors, while using a supporter tier as a lightweight way for sympathisers to express commitment and reduce funding concentration risk.
If I were to bet on what’s happened here, I’d bet it’s something to do with Thomas leaving.
Looking at his LinkedIn and his Forum history, he seems very well connected in the field of AI safety.
I suspect it was easy to get funding because people knew and trusted him.
EA community building relies heavily on a few large donors. This creates risk.
One way to reduce that risk is to broaden the funding base. Membership models might help.[1]
Many people assume EA will only ever appeal to a small slice of the population, and so this funding would never amount to anything significant. However, I think people often underestimate how large a “small slice” can be.
Take the Dutch mountaineering association. A mountaineering club in one of the flattest countries on Earth doesn’t exactly scream mass appeal.
So, how many members do you think it has?
Around 80,000. That alone brings in roughly €4 million in membership contributions—about 70% of its total annual income.[2]
Even niche communities can fund themselves if enough people are engaged.
Congrats on your first Forum post—really glad you shared it.
One quick suggestion: I’d frame this more as an intervention idea than a cause area. It seems to fall under broader categories like “improving human wellbeing” or “global mental health”.
I’d also question the likely cost-effectiveness of interventions here, at least in the near term. As you note, the problem is most acute in high-income countries, places where the marginal impact of philanthropic interventions tends to be lower. That said, as smartphone access grows and incomes rise in the Global South, these harms could become much more widespread, so this may well be a growing global issue over time.
You might find this post on violence against women and girls a helpful reference. It’s another example of someone exploring a neglected social issue and trying to assess it in EA terms (scale, neglectedness, cost-effectiveness, etc.). I think a similarly structured dive into the screen addiction issue—perhaps comparing potential interventions and DALY burden—could be a great next step.
Thanks again for posting and I hope you keep going!
I think you’re right about the limitations of these examples, but this feels like we’re getting lost in the weeds. The original point was about travel costs making democratic decision-making processes suboptimal in large countries. These examples show that’s not true—organisations routinely manage democratic processes across large geographies.