Beyond AI hype: building new institutions for AI and the Digital Revolution

The bingo card that forgot AI

In hindsight, it’s odd. I spend a good chunk of my waking hours thinking, writing, and talking about data, AI and the governance thereof, and yet my 2026 “existential dread" bingo card back in January was full of climate risk, geopolitical tensions, pandemics, and domestic politics - and AI was nowhere to be seen.

It’s not that AI wasn’t on my radar. By 2026, we have already had successive waves of generative AI hype, major advances in multimodal systems, the emergence of agentic AI, and a steady emergence of new AI governance initiatives from the EU AI Act to the AI safety summits in the UK and South Korea. AI was clearly “a thing.” But it didn’t make the bingo card.

Reflecting on that now, I think it’s because AI, for me, is feeling less like a discrete technology risk and more like an epochal shift - something closer to a Digital Revolution on the scale of the Industrial Revolution. And once you start thinking in those terms, the bingo-card frame begins to feel a bit small.

AI as a digital revolution

Historians and economists have already begun making the comparison between AI and the Industrial Revolution, not as a lazy metaphor but as a serious analytical exercise. The Industrial Revolution fundamentally altered how societies produced value: steam and mechanisation transformed physical labour, reorganised cities, and reshaped class structures.

AI is doing something analogous for cognitive and organisational work. Instead of steam engines and power looms, we have large language models and recommendation systems augmenting or automating tasks that once required human judgment - from document review and coding, to medical image analysis and financial modelling.

Some recent research suggests that AI and big data technologies are already shifting the labour share of income in ways that look eerily similar to the Industrial Revolution, with early evidence of a 5-15% decline in labour’s share in some sectors. At the same time, firms adopting AI are hiring more, not fewer, people - particularly those with AI skills - which echoes the historical pattern where new technologies destroyed some jobs but created entire new categories of work.

You can dive into some of these analyses here:

• Columbia Business School’s research brief on AI and labour share, which explicitly compares AI’s economic impact to the Industrial Revolution.
• A 2026 VoxDev piece unpacking similarities and differences between AI and the Industrial Revolution, including lessons for managing inequality and social disruption.

The upshot is that if AI really is an Industrial-Revolution-scale transformation, we should expect not just productivity gains and efficiency, but also intense distributional conflict, institutional stress, and new forms of social organisation.

Which brings me back to a much earlier period of my life, long before “AI and data governance” was a job description.

A minor reminscence: the English working class, 1780–1945

When I was an undergrad studying history at the University of Sydney (a very long time ago), one of the units that stayed with me was on the English working class from 1780-1945. We looked at that long arc of upheaval - enclosure, industrialisation, urbanisation, war, depression - and how people, especially workers, carved out spaces of agency and dignity within it.

One of the most striking aspects of that period was not just the exploitation and hardship, but the incredible institutional creativity of the working class. In the midst of industrial capitalism’s dislocations, people built new and significant institutions:

• Working men’s clubs: Emerging from the mid-19th century, they were designed as places where working men could gather for “conversation, business and mental improvement, with the means of recreation and refreshment.” The Club and Institute Union, founded in 1862, explicitly aimed to support clubs and institutes that combined social life with education and self-improvement. Many were built around “harmless amusements such as chess” combined with practical education and penny savings banks, giving workers both social connection and financial tools.
• Schools of the arts and mechanics’ institutes: Across Britain (and in Australia), these institutions offered lectures, libraries, and classes aimed at adult education, especially in technical and scientific subjects that were relevant to industrial work.
• Friendly Societies: These mutual aid groups, like the Oddfellows and Foresters, formed from the late 18th century to provide sickness benefits, funerals, and insurance for members excluded from early welfare systems. They emphasized self-help and community solidarity amid harsh factory conditions.
• Trade Unions: Early combinations evolved into formal unions by the 19th century, such as the Grand National Consolidated Trades Union (1834), advocating for better wages and hours. They served as protective networks and political voices for workers.
• Co-operative Societies: Inspired by Robert Owen, groups like the Rochdale Pioneers (1844) created stores and workshops for fair pricing and worker ownership, blending economic mutualism with community hubs.
• Chartist and Mutual Improvement Societies: Chartist groups (1830s–1850s) and lyceums offered lectures, libraries, and debating clubs for political education and self-improvement, fostering working-class activism. Sunday schools also doubled as literacy centers.

Historians like Ruth Cherrington have documented how these clubs and institutes offered a blend of recreation, political discussion, and education, often deliberately avoiding being too educational so as not to alienate their members. Others, like T. G. Ashplant, emphasise that they were attempts to create spaces where working people could participate in “mental improvement” on their own terms.

These were not perfect institutions; they were contested, often paternalistic, and many times exclusionary. But they were also genuine innovations in social infrastructure, emerging from and for a class that was being rapidly reshaped by industrial capitalism. They provided venues for mutual aid, political organising, cultural life, and informal learning - all under conditions of intense technological and economic disruption.

What kinds of institutions does an AI age need?

Fast forward to 2026, and we have our own landscape of AI-related institutions:

• Global bodies like the UN, OECD, and UNESCO setting high-level AI principles and recommendations.
• Standards organisations such as ISO, IEC, IEEE, and NIST proposing technical and governance frameworks for trustworthy AI.
• National regulators implementing instruments like the EU AI Act, competition interventions, data protection enforcement, and sector-specific AI guidance.
• A growing ecosystem of research centres, ethics boards, and governance alliances working on responsible AI.

If you want a sense of the sheer density of this emerging governance layer, check out this global AI governance map that tracks over 140 institutions across four layers: global governance, standards, regulators, and ethics/oversight bodies.

These are all important - and, in some cases, overdue. But they are mostly top‑down or expert‑centric institutions: treaties, standards, regulatory agencies, specialist research centres. They occupy a necessary slice of the governance stack, but they don’t quite answer the question that my undergraduate history unit has left me with:

Where are the AI-era equivalents of these working men’s clubs and schools of the arts?

In other words:

• What institutions are we building that allow ordinary people - not just policymakers, engineers, and CEOs - to collectively shape how AI shows up in their work, neighbourhoods, and lives?
• Where do people go to experiment, learn, and organise around AI, beyond being passive recipients of platforms and products?
• How do we create spaces that feel as normal and accessible as the local club or community hall, but that are explicitly designed for navigating the AI transition?

Interestingly, there are early proposals that nod in this direction at the global level: ideas like a “CERN for AI” or a “Global AGI Agency” that would be public-private partnerships to develop and govern advanced AI systems as a kind of global public good. The Future of Life Institute, for example, has solicited designs for new global institutions, ranging from Fair Trade AI schemes to multilateral AI agencies embedded in both the UN system and industry.

These are important experiments. But they are still far from the lived reality of most communities. They also risk being primarily elite projects, even if well-intentioned.

Imagining AI-era social infrastructure

If we take the Industrial Revolution analogy seriously, then institutional innovation shouldn’t be an afterthought - it should be central. Just as the 19th-century working class built clubs, institutes, co‑operatives, unions, and friendly societies, perhaps we need a wave of AI-era social infrastructure that is just as inventive.

Some possibilities that intrigue me:

• AI literacy and practice hubs: Think of a contemporary School of the Arts but oriented around digital skills and AI literacy - not just coding, but critical understanding of data, models, labour impacts, and governance. These could be run through libraries, TAFEs, universities, or community organisations, with a mix of formal and informal learning.
• Community AI labs: Small, local spaces where people can experiment with AI tools on their own terms - building things that matter to them (local language models, mutual aid tools, civic data projects) rather than just consuming whatever Silicon Valley ships next.
• Worker-led AI councils: Sectoral or workplace-based bodies where workers, unions, and professional associations co-design how AI is adopted, including rules about surveillance, deskilling, and benefit-sharing. This echoes some of the way working men’s clubs and institutes became spaces for political as well as educational activity.
• Civic observatories for AI impacts: Local or regional institutions that track AI’s effects on employment, housing, education, health, and social cohesion - a kind of “weather service” for AI’s social impacts - and feed that data back into policy debates.

None of these ideas are fully fleshed out, and many would bump into resource constraints, power imbalances, and political resistance. But that was also true in the 19^th century. The institutions that endured were not the product of a single master plan; they were the cumulative result of many experiments, some of which failed, some of which evolved, and some of which were co‑opted.

Beyond dread: putting AI on the bingo card differently

So what do I do with all this when I sit down next January to sketch out my 2027 existential dread bingo?

AI probably will appear on the card next time, but not just as “runaway AI risk” or “AGI apocalypse.” Instead, I suspect it will show up as something like:

• “We fail to build the institutions this transformation requires.”
• “We leave AI governance to a narrow set of actors and miss the chance for broader democratic input.”
• “We treat AI purely as a productivity tool, not as a catalyst for rethinking how we live, learn, and work together.”

The Industrial Revolution analogy is useful not because history repeats itself mechanically, but because it reminds us that technology shifts of this magnitude always come with institutional and social upheaval. We can’t avoid that, but we can shape it - we still have agency as human beings.

Back in that undergraduate history unit, I learned that the working class didn’t just endure the Industrial Revolution; they responded to it and reshaped it, in part by building their own institutions. As we navigate the AI upheaval, perhaps the most important question is not “What will AI do to us?” but “What will we build in response?”

And maybe the bingo card is not a bad place to start - not just to catalogue our dreads, but to surface the institutions that we still need to imagine into existence.