As artificial intelligence reshapes economies faster than most governments can regulate it, a new kind of civic institution has sprung up to fill the gap: the AI meetup. Organisations like AI Salon and the AI Collective, both born in San Francisco within months of ChatGPT’s launch, now host hundreds of local chapters worldwide. Their aim is to give space for people to discuss what AI means for their jobs, their communities, even on their continents—without waiting for regulators or corporations to fill in the blanks.
The pitch is appealing. At RESET, we’ve made no secret of our support of grassroot movements for civil good, especially in the tech industry. Especially given not only the AI industry’s tendency toward bullish business practices and that just a small cluster of American firms control the technology’s direction, these groups offer something regulators and companies can’t: a local, peer-led space to make sense of the disruption.
But their own founders, with the best intentions or not, are wrestling with a harder question: can organisations rooted in Silicon Valley culture build truly helpful outcomes for those who seek to improve their own local situations? Or, do they end up reproducing the very hierarchies they were formed to counter instead?
From a living room to a movement
AI Salon started small: a single conversation in a friend’s apartment, organised by neuroscientist and AI safety researcher Ian Eisenberg. Since then, he has run roughly 40 salons, each capped at about 20 people. Their premise is simple: keep an open mind, and everyone speaks at least once. Sessions are generally themed across AI’s intersection with climate, education, consciousness and art. The nonprofit has since spread beyond San Francisco to volunteer-run chapters in everywhere from Berlin and Bengaluru to Singapore and Sydney.
Eisenberg’s motivation was partly a reaction to the tech industry’s usual gathering places. “There are a lot of hackathons and startup meetups, but they are excluding a lot of people, and we want broader civil society participation,” he told Rest of World in August 2026. He added that with public trust in governments on the decline, people are increasingly left to “figure out” the implications of AI on their own.
The AI Collective has scaled that idea further. Founded the same year, it now counts more than 250,000 members across roughly 200 chapters in 50 countries, run mostly by volunteers alongside a small paid core team. Its stated mission is to “democratise, build trust, and coordinate how society navigates the rapid acceleration of technological progress”.
Chief Operating Officer Catherine McMillan says the long-term goal is for membership to tip toward being mainly non-American; a deliberate move away from the organisation’s Silicon Valley centre of gravity. “We want to encourage hyper-locality, building the most local communities around AI,” she told Rest of World.
A practical edge
Talk is one thing. But a central question is: how useful are these meetups if nothing happens afterwards? For the AI Collective in particular, a growing calendar of event programming goes beyond discussion toward something more hands-on.
Recent sessions have included Practical AI at Work: Turn Your Brand Guide Into an AI-Ready Skill, run over Zoom for members looking to apply the technology directly to their own small businesses, and Technology, AI and Addictive Behaviors: Can They Be Treated?, which brought in researchers and clinicians working at the front line of digital-addiction treatment to unpack how social platforms, gaming and AI systems are reshaping attention, emotional regulation and everyday behaviour.
However, turning a brand guide into a shiny, AI-modified product is a different kind of problem than the ones facing members in Nairobi or Bengaluru. There, the pressing questions needing practical input aren’t necessarily about optimising a workflow, nor how to get over an addiction to a technology they may not even have access to.
Where the real test lies
That meetup concept is being tested hardest in places where AI adoption is still nascent and the stakes of getting left behind are highest. In Nairobi, where AI adoption sits below 10 percent—among the lowest rates globally—the AI Collective’s 3,000-member chapter is trying to make sure the technology’s benefits in education, agriculture and healthcare don’t bypass the country entirely.
This is on top of addressing members’ concerns about the technology’s impact on data privacy and job security. Organiser Francis Kangure frames the chapter’s role as “a bridge between local talent and global opportunities” and a response to what he calls “a huge knowledge gap regarding AI and AI adoption in Africa”.
In Bengaluru, the concern is subtler but no less pointed. India is already Anthropic’s second-largest market for Claude. Yet, chapter head Abhishek Kumar says the country’s AI story is being written top-down, by national policy and large corporates, with little room for individuals. “There’s a quiet worry about whether India ends up only consuming and wrapping models built elsewhere, instead of building foundational technology of its own” he explains. He also points to the issue of whether people who don’t speak English get left out of the picture altogether.
That worry isn’t hypothetical. AI systems remain overwhelmingly built for English, Mandarin and a handful of other commercially valuable languages, leaving thousands of other language communities effectively excluded from the technology’s benefits. This gap has real, tangible effects. In rural Indonesia, for example, health IT systems built without adequate translation into Javanese and Sundanese produced dangerous errors in medication dosages.
Likewise, for much of the Global South, where the impacts of AI’s rapid development are being felt the hardest, a lack of access to AI isn’t just an inconvenience but actually carries real stakes for self-determination and development.
AI’s lack of linguistic diversity marginalises millions worldwide, compounding economic, social and technological inequalities. However, efforts to create a more inclusive AI future are underway.
Read more: How AI’s Failure on Linguistic Diversity is Deepening Global Inequality
Inclusion isn’t the same as power
It’s here that the meetup model runs into a sharp critique. We asked Uchechukwu Ajuzieogu, Founder and Editor-in-Chief of Aylgorith and Executive Director of the African Institute for Artificial Intelligence Policy, about what he calls multi-stakeholder “AI democratisation”, something he’s researched at length. He mentioned Nigeria’s own version of the AI meetup—the Nigeria AI Collective—backed by Luminate, the Ford Foundation and AI Collaborative. When we asked him whether meetups like the ones mentioned have come up in his research of improving pay, working conditions or bargaining power for those involved in AI’s development, his conclusion is blunt: “Short version: not one”.
What actually moved wages and conditions for African data workers, he argues, came from somewhere else entirely: unionisation, litigation and independent audits.
He points to Nairobi, where more than 150 AI content moderators formed the African Content Moderators Union in 2023. That organising, combined with strategic litigation led by lawyer Mercy Mutemi and the NGO Foxglove, got Kenyan courts to rule that Meta could be sued directly, to block mass layoffs and order back pay. Separately, sustained pressure from Fairwork’s labour audits pushed Sama, the outsourcing firm behind Facebook’s moderation work, to substantially raise its lowest-paid moderators’ wages. “None of that came from a meetup” says Ajuzieogu. “It came from workers organising, lawyers filing suit and auditors publishing scorecards that named the company.”
Data labour powering generative AI are paid shockingly low wages, making them particularly attractive to AI service providers who seek to maximise profit. But, there’s hope.
Read more about Uchechukwu Ajuzieogu’s work on revealing the hidden plight of data workers in this article from 2025: “Magic” AI Is Exploiting Data Labour in the Global South – But Resistance Is Happening
“A meetup can be where people find each other”
He frames the gap using a term increasingly common in the academic literature: participation-washing. Researchers studying AI governance forums distinguish “inclusion”—being consulted, attending or having a seat—from “power”, actual decision-making authority. Multi-stakeholder groups, in his reading, tend to deliver the first and not the second. As he put it, the real test of these forums isn’t how many stakeholders were in the room, but whether participation redistributed power as opposed to simply giving dissent somewhere polite to go.
However, Ajuzieogu doesn’t dismiss the meetups outright. “A meetup can be where people find each other, that has some value”, but he’s wary of collapsing that value into something bigger than it is. Treating the meetup itself as the mechanism of change, he warned, “substitutes the feeling of progress for the thing itself”.
The unresolved tension
This is where the promise of these grassroots collectives runs up against their own origin story, as well as evidence provided by researchers such as Ajuzieogu. Groups like AI Salon and AI Collective that are founded in Silicon Valley, staffed (at least initially) by people fluent in its culture and structured around its networks, are trying to become genuinely global institutions that serve communities the industry has historically ignored. This deserves some praise. The chapters in Nairobi and Bengaluru show these organisations can, at their best, highlight real local needs such as language and infrastructure gaps.
But highlighting a problem isn’t the same as solving it. And perhaps the two are being conflated more often than the organisers might like to admit. Whether the AI Collective and groups like it can actually shift power outward, rather than simply extending Silicon Valley’s reach through a more diffuse, friendlier network of volunteers, will depend on whether local chapters gain real influence over these organisations’ direction, and not just a seat in the room. It may also depend on whether they’re honest about what a meetup can and can’t do: a place to find allies, perhaps, but not, on the evidence so far, a substitute for a union, a lawsuit or an audit with teeth.
