The Phrase That Won't Go Away
"Fourth Industrial Revolution" is everywhere. Every tech conference keynote. Every consulting firm's thought leadership report. Every breathless LinkedIn post. The World Economic Forum's been pushing the term since 2016. By now it's become one of those phrases that's so overused it's almost lost meaning.
So let's actually think about this. What were the first three? Is AI really number four? And does the label even matter?
The First Three, In Brief
The first industrial revolution kicked off in Britain around the 1760s. Steam power, mechanized textile production, iron-making improvements. A weaver in 1750 produced cloth by hand. A weaver in 1800 operated a power loom that did the work of twenty people. That's the fundamental pattern of an industrial revolution — a technology that multiplies human productive capacity by an order of magnitude.
The second came roughly a century later, built on electricity, steel, and the assembly line. If the first revolution was about mechanization, the second was about mass production. The Model T Ford came off a moving assembly line in 1913. By 1927, Ford had made 15 million of them. That kind of scale would have been unimaginable to a craftsman in 1850.
The third revolution started in the 1960s and 1970s with semiconductors, personal computers, and eventually the internet. This one was about information. A factory worker in 1950 controlled one machine. A knowledge worker in 2000 could coordinate supply chains spanning continents, communicate with anyone instantly, and access the sum of human knowledge from a device in their pocket.
Each revolution didn't just change how things were made — it changed how people lived, where they lived, what skills were valuable, and who held power.
The Case For AI As The Fourth
The argument for AI as the fourth industrial revolution goes something like this: for the first time, we're automating cognition rather than physical labor.
The steam engine automated muscle power. The computer automated calculation and information storage. AI automates judgment, pattern recognition, and creative generation — things that, until very recently, we assumed required human intelligence.
If you can automate judgment at scale, the implications are enormous. A lawyer in 2026 who uses AI for document review might handle ten times the caseload of a lawyer in 2016. A radiologist using AI-assisted diagnosis might catch cancers that would have been missed. A software developer with AI tools might produce working code 2-5x faster than without them.
The scale of potential economic impact is genuine. Goldman Sachs estimated in 2023 that generative AI could increase global GDP by 7% over ten years. McKinsey put the number even higher. Even if you discount consultant estimates by half — which is usually wise — we're still talking about trillions of dollars in economic value.
The speed of adoption is also historically unusual. ChatGPT reached 100 million users in two months. The telephone took 75 years to reach that many users. The internet took about seven years. When a technology spreads that fast through the global economy, it's reasonable to wonder whether we're seeing something fundamental shift.
The Case Against
Okay, counterarguments.
First, calling something a "revolution" before it's actually revolutionized anything is premature. The first industrial revolution wasn't obvious to the people living through it. Most people in 1790 were still farmers whose daily lives hadn't changed much. The revolution is only visible in retrospect. We might be calling AI a revolution because it's exciting right now, not because we know how it plays out.
Second, a lot of the economic impact projections assume that AI capabilities will continue improving at their current rate, and that the improvements will translate directly into productivity gains. Neither assumption is guaranteed. There could be diminishing returns to bigger models. The hard problems of reliability, factual accuracy, and reasoning might prove more stubborn than the scaling curves suggest. The gap between "AI that impresses in a demo" and "AI that you can trust with actual work" could remain wide.
Third, previous industrial revolutions created entirely new categories of work. The first revolution didn't just make existing farmers more productive — it created factory jobs, engineering disciplines, and eventually an entirely new economic structure. Is AI creating genuinely new categories of economic activity, or mostly automating existing ones? I'm not sure yet.
Fourth, the productivity paradox. Remember how computers were supposed to dramatically increase productivity? In 1987, Robert Solow famously quipped, "You can see the computer age everywhere but in the productivity statistics." Productivity growth in the US actually slowed during the early decades of computer adoption. It wasn't until the late 1990s that computers started showing up as measurable productivity gains. We might be in a similar period with AI — seeing the technology everywhere but not yet seeing it in the numbers.
What's Actually Different This Time
So here's my attempt at a balanced take.
AI in 2026 is not just hype. The capabilities are real, they're improving quickly, and they're already changing how work gets done in specific industries. Software development, content creation, customer service, legal document review — these fields have already been meaningfully affected.
But the "fourth industrial revolution" framing implies a clean break — before and after. I think that's wrong. What we're seeing is more like an acceleration of the third revolution rather than the start of something entirely new. AI is the logical extension of digitization, not a departure from it. The computer revolution put information at our fingertips. AI makes sense of that information and acts on it.
The thing I'm most confident about: the biggest effects of AI won't be obvious for another decade. The first industrial revolution took fifty years to really transform society. The notion that AI will remake the economy in five years is a fantasy. The notion that it won't change anything fundamental is probably also wrong. The boring truth is somewhere in the middle — gradual, uneven, more evolution than revolution, but evolution at a pace we haven't seen before.
What I'm Watching
A few things will tell us whether this is genuinely revolutionary or just an unusually fast technological cycle:
Whether AI creates new economic sectors rather than just making existing ones more efficient. The internet gave us social media, search engines, e-commerce, the app economy — categories that didn't exist before. What are the AI-native industries that don't map to anything we had before?
Whether the productivity gains show up in national statistics. If AI is as transformative as its advocates claim, we should see it in GDP numbers and productivity data within five years. If we don't — Solow's paradox, round two.
Whether the benefits distribute broadly or concentrate narrowly. The first industrial revolution created enormous wealth but also enormous inequality and suffering for the workers who lived through the transition. The question isn't just "will AI make us richer on average" but "who gets richer and who gets left behind."
I don't have answers. I don't think anyone does yet. But those are the right questions.
