May 1st, 2026: The Work We Inherited

TL;DR:

25% of global employment is exposed to significant changes from generative AI (ILO, 2025).
Every industrial revolution destroyed jobs massively... and then created many more. But the transition period was always painful and unequal.
This time, white-collar workers are the most exposed — developers, lawyers, journalists — not manual laborers.
Adaptation is possible. It's not easy. But it's possible.

Today is May 1st. Workers' Day.

And I'm not sure how to celebrate something that, in 2026, feels strange to even name.

Because as I write this, headlines are full of mass layoffs in Silicon Valley, startups that replaced half their team with AI agents, consulting firms automating tasks that two years ago were the core work of people with university degrees. And in Argentina, in Mexico, in Spain, the conversation is the same: how long do I have before something — or someone — does what I do, faster and cheaper?

It's a scary question to ask out loud. So today, I'm asking it.

This isn't the first time work has broken down

Before we talk about what's happening now, we need to zoom out. Because what we're living through today isn't the first time a technology shook the foundations of how we work. Not the second. Not the third.

1760. The First Industrial Revolution. The steam engine arrives. Artisans who spent generations weaving cloth by hand watch factories displace them. Child labor explodes. Cities fill with people migrating from the countryside, not knowing quite where they're headed. Conditions are brutal: 16-hour shifts, no safety, no rights. Unions are still illegal in Britain.

The result? Massive destruction of artisan jobs. And the creation, over time, of an industrial working class that had never existed before.

1870. The Second Industrial Revolution. Electricity arrives, then the combustion engine, the telephone. Ford introduces the assembly line. Work is fragmented into specialized, repetitive tasks. Large corporations grow. Migrations become massive — toward industrial cities in Europe, toward Argentina, toward the United States.

The result? Displacement of millions of agricultural and craft workers. And the consolidation, also over time, of the labor movement, labor legislation, the 8-hour workday — precisely what we celebrate every May 1st.

1970. The Third Industrial Revolution. Computers automate manufacturing. Assembly work that used to be done by humans becomes robotic. Industrialized countries begin to deindustrialize. The services sector grows. Precariousness appears: temporary contracts, outsourcing, work fragmented into projects.

The result? A collapse in manufacturing employment in the West. And the rise of the knowledge economy — new jobs that didn't exist, for tasks that didn't exist, in industries that didn't exist.

Do you see the pattern?

Every revolution destroys massively. Every revolution also creates massively. But the transition period is always painful, unequal, and doesn't wait for anyone.

What's happening right now

The difference with this Fourth Revolution — AI, language models, cognitive automation — is that this time it's not manual or repetitive jobs in the crosshairs.

It's ours.

According to the International Labour Organization (ILO), 25% of global employment is in occupations with some degree of exposure to generative AI. The most affected are administrative and desk tasks: analysis, writing, support, standard coding, legal review, routine accounting.

In other words: the person most at risk today is not the plumber. It's the junior developer. The executive assistant. The content writer. The financial analyst. The lawyer who reviews standard contracts.

And there's already data on what's happening in practice. In Spain, job offers requiring AI skills grew 680% between 2018 and 2024. In Latin America, AI specialist salaries reach $7,000 per month in countries where the average wage is around $800. Companies integrating AI report productivity jumps of 14% to 70% in specific tasks.

But on the other side: companies that laid off programmers trusting automated coders, then had to hire human engineers to fix what the AI broke. A new job category is emerging that nobody predicted: "AI repairers" — people hired to fix automatically generated texts, images, or code. Often paid less than the original position that was eliminated.

It's not simple. It's not linear. And it's not equitable.

Who's most exposed?

This is what strikes me most about the data: AI doesn't affect everyone equally.

Women in high-income countries have greater risk exposure than men in the same markets. 9.6% of high-risk female employment in wealthy countries versus 3.5% for men, according to ILO reports. Part of the explanation: the most automatable jobs in this cycle are administrative and support roles, where historically more female employment is concentrated.

The urban-rural gap is also widening: workers in information-intensive sectors (finance, technology, professional services) have more exposure than those in sectors with a stronger physical or manual component. In Latin America, the digital talent crisis is critical: the supply of professionals with AI skills is insufficient, but demand is growing in roles that aren't even technical — sales, marketing, support.

And there's salary polarization: workers with AI skills earn up to 56% more than their peers without those skills in the North American market. The gap isn't between "those who work" and "those who don't." It's between those who updated themselves and those who couldn't — or didn't have the means.

How to adapt without being consumed by fear

Here I need to be honest, because there are two traps that are easy to fall into.

The first trap is empty optimism: "don't worry, new jobs always emerged, this will pass too." Technically true. Historically accurate. But that doesn't say anything to the 45-year-old who spent 20 years in accounting and just watched their main task get automated. Transitions are real. The pain is real. And "everything will be fine in the long run" doesn't pay next month's rent.

The second trap is paralyzing panic: "AI is going to replace all of us, there's nothing we can do." That's not true either. The ILO is clear about this: jobs are more likely to transform than disappear. The example I liked most from the research I read is a KPMG auditor who developed an AI agent to process reports — which freed her time to do the risk analysis and professional judgment work that actually matters. It didn't replace her. It gave her time back for the genuinely human part of her work.

Adaptation has layers. And not all the responsibility falls on individuals.

For those of us currently in the labor market:

Continuous learning stopped being a plus and became the only strategy. You don't need to become an AI engineer — you need to understand how to use AI tools as assistants in your own work. A lawyer who knows how to work with language models for contract review isn't replaced by AI. They're the lawyer who can do the work of five in the time of one.

The skills that will be most valued in the coming years are the ones AI can't replicate well yet: complex critical thinking, non-routine creativity, empathy, leadership, negotiation, ethical judgment. These aren't soft skills. They're hard skills that happen to be human.

For companies:

The case of companies that laid off programmers then re-hired them to fix what the AI broke is the best possible argument for investing in internal training. AI works better as a tool than as a substitute. The organizations winning in this transition aren't the ones that replaced people with algorithms — they're the ones that trained their teams to work with AI and supervise it.

For governments and public policy:

Singapore launched a national program to train 100,000 "AI-bilingual workers" by 2029. It's not perfect, but it's a signal of what's needed: public training infrastructure, digital access for rural areas, tripartite dialogue (government-companies-unions) to regulate transitions, and safety nets that don't leave out platform workers and the gig economy.

May 1st in 2026

Today, May 1st, I'm not celebrating the arrival of a technological utopia where nobody works because AI does everything. Nor am I joining the apocalypse that human work has ended.

What I'm celebrating — and defending — is the idea that work must remain human at its core. That the productivity AI generates must be distributed fairly, and not just accumulate in the hands of those who already have the most. That the rights won by generations of workers — the 8-hour day, social security, the right to organize — cannot be left out of the gig economy, platform contracts, or new forms of digital work.

History shows that industrial revolutions don't respect anyone who stays still. But it also shows that revolutions don't have to be unjust — if there's political will, collective organization, and a decision not to leave anyone behind.

Every May 1st is a reminder that work is not just economics. It's identity, community, dignity.

And that, for now, no AI can generate.

References

Cris Fernandez — Social intelligence for the world ahead 💙 The work of the future is being built today. But only if we understand what's happening.