If you've spent any time on social media lately, you've probably heard some version of this: "You have five years left to get rich. After that, AI takes over, and whoever owns the machines owns everything." It usually comes with urgent music and a link to a course.

It's a great hook. It's also a great exercise in futures thinking, because it bundles a real signal with a made-up deadline. So let's take it apart the way we would any other forecast: separate the signal from the noise, check the timeline against history, and see what's left.

Step one: find the signal

Futures thinking starts by asking what's actually changing, before asking how fast.

And here the claim isn't pulling from nowhere. There's serious research suggesting AI could shift wealth toward people who own capital. An IMF working paper from April 2025, AI Adoption and Inequality by Emma Rockall, Marina Mendes Tavares and Carlo Pizzinelli, modeled this using UK household data. Their result is counterintuitive: AI could actually reduce wage inequality, because it mostly affects well-paid cognitive jobs. But it could increase wealth inequality, because returns on capital rise and the people who already hold assets benefit most.

So the direction behind the viral claim, "AI may favor owners over workers," has real support. That's the signal. Everything else is packaging.

Step two: check the timeline

This is where the claim falls apart. Nothing in the serious forecasts points to a five-year window that closes.

Economists don't even agree on how big AI's impact will be. Daron Acemoglu (MIT, Nobel laureate in economics in 2024) estimated in 2024 that only about 5% of tasks would be profitably automated within ten years, adding a "nontrivial but modest" 1% or so to GDP over that decade. Goldman Sachs had earlier projected something closer to a 7% boost to global GDP over ten years. That's a big range, and neither number describes a world that flips in five years.

History suggests patience too. In 1987, economist Robert Solow joked that "you can see the computer age everywhere but in the productivity statistics." It took until the mid-1990s for computers to show up clearly in productivity data. Erik Brynjolfsson and colleagues describe this pattern as a "productivity J-curve": when a new general-purpose technology arrives, measured productivity can dip at first while companies reorganize, retrain and build the complementary pieces they need. The payoff comes later.

A useful rule from futures work: be suspicious of any forecast that is precise about timing and vague about mechanism. "Five years" is very precise. "AI takes over" is very vague.

Step three: look for historical patterns

One of the best tools for thinking about technological change and money is Carlota Perez's Technological Revolutions and Financial Capital (2002). Perez describes five great surges since the Industrial Revolution (starting with the factory system in 1771, then steam and railways, steel and electricity, oil and the automobile, and information technology). Each followed a similar arc:

1. Installation. The new technology arrives and financial capital rushes in. This phase often ends in speculative frenzy, bubbles and rising inequality.
2. Turning point. Usually a crash, followed by new rules and institutions.
3. Deployment. The technology spreads through the whole economy, and gains are shared much more widely.

Each full surge took roughly half a century.

Look at the viral claim through that lens. "Get in now before the window closes" is exactly the kind of story that circulates during a frenzy. In Perez's model, the frenzy does make some people very rich, mostly those who already had capital or were early and lucky. It also tends to end in a crash that wipes out a lot of latecomers. And the broad wealth creation comes after, during deployment.

Whether AI is a new surge or the late stage of the information revolution is still debated. Either way, the pattern points the opposite way from the claim. If there's a window, it isn't closing in five years. The bigger changes are more likely to unfold over decades.

Researchers at the Santa Fe Institute offer an even longer view. Studying ancient societies, they found that wealth inequality rose sharply after the ox-drawn plow made land and animals the key assets. But it happened over many centuries, and political and cultural choices shaped how far it went. Technology creates the conditions. People and institutions decide the outcome.

Step four: map the futures, not the future

Futures thinkers often use the "futures cone," popularized by Joseph Voros: a way of sorting futures into possible, plausible, probable and preferable. It helps to put the viral claim into that cone.

Probable: AI keeps spreading unevenly across sectors and countries. Owners of productive assets benefit disproportionately. Change is gradual enough that 2030 looks more like "today, but more so" than a different world. The World Economic Forum's Future of Jobs Report 2025, for example, expects about 170 million jobs created and 92 million displaced by 2030: a lot of churn, but not the end of work.

Plausible: an AI-related investment bubble bursts before the technology's benefits spread widely, as has happened in previous technological surges. Or winner-take-most dynamics concentrate gains in a handful of companies faster than in past cycles.

Possible but unlikely: a hard, permanent line in 2030 after which ordinary people can never build wealth again. No serious model I found supports that.

Preferable: this is the part the viral videos skip. Outcomes depend heavily on policy: taxes, education, labor rules, public investment. The same IMF paper shows how much policy choices change the distribution of gains. The future isn't just something that happens to us.

What to do instead of panicking

If you strip away the countdown, the useful advice is boring, and that's a good sign.

• Build slowly and diversify. Deployment, where broad gains happen, is still ahead. Steady, diversified saving has historically beaten trying to catch a closing window.
• Invest in complementary skills. The same IMF work suggests workers whose tasks complement AI fare better than those whose tasks it simply replaces. For designers, that means judgment, research, facilitation, taste and the ability to direct tools, not just operate them.
• Watch for weak signals. Futures work is about noticing early signs: shifts in investment, regulation, hiring patterns, what your own clients start asking for. They'll tell you more than any countdown.
• Prepare for several futures. Don't bet everything on one story, especially one sold to you with a deadline.

A small exercise

Next time you see a confident prediction about AI, run it through three questions:

1. What's the signal underneath, and is there real evidence for it?
2. Where does the specific timeline come from?
3. Who benefits if I believe it right now?

The five-year claim fails the second question and, honestly, the third one too. But it passes the first, and that part is worth paying attention to.