A lot of old predictions about 2025 missed the details but nailed the direction: more connected lives, more AI in everyday systems, and bigger debates over who benefits. That matters now because many 2035 forecasts risk making the same mistake — overestimating the gadget, underestimating the social fallout.

If you go back and read older forecasts about “the world of 2025,” the first thing that stands out is not how wrong they were. It’s how often they were broadly right in ways that looked vague at the time.

In 2014, Pew Research Center’s “Digital Life in 2025” project described a future where connectivity would become less visible and more embedded — “like electricity” — with AI-assisted systems woven into daily life. That now reads less like sci-fi and more like a description of ordinary life: always-on phones, cloud services, algorithmic feeds, recommendation engines, smart devices, and AI tools layered into search, work and communication.

That doesn’t mean the forecasts were precise. They usually weren’t. They were often better at identifying direction than timing, adoption speed, and who would actually get access.

The internet is a good example. The International Telecommunication Union said its 2025 estimates show roughly three-quarters of the world’s population online — about 6 billion people — while 2.2 billion people remain offline. So the “ubiquitous connectivity” prediction was partly right, but the universal version of it still hasn’t arrived, and the gap is not small.

The same goes for AI. Pew’s 2014 expert canvassing on AI and robotics expected major advances and deeper integration into ordinary life by 2025. That basic call aged well. AI is now embedded across consumer and business software, often as a feature rather than a standalone product — including writing assistants and workflow tools built directly into the office software many people use every day. But many older forecasts treated adoption as mostly a technical question. In practice, the friction has been social and institutional too: skills gaps, trust, regulation, workplace redesign and uneven access.

Some forecasts also overshot how quickly “breakthrough” technologies would become normal. The World Economic Forum’s 2020 collection of expert predictions for 2025 included strong claims about areas like quantum computing, health system shifts and 5G-enabled transformation. Parts of that have moved forward, but not evenly, and not at the same pace across sectors or countries.

What’s useful here is not dunking on old predictions. It’s noticing what they consistently miss.

They tend to underestimate bottlenecks: infrastructure build-out, affordability, standards, politics, workforce training, and public trust. They also tend to underestimate how long “messy middle” periods last — the years when a technology is real, important, and still not delivering the clean future people imagined.

That matters for 2035, because today’s forecasts are already repeating the pattern.

The big themes in current forward-looking reports — AI-driven labour market shifts, demographic change, energy transition, and fragmentation in the global economy — are likely to shape the next decade. The World Economic Forum’s Future of Jobs Report 2025 frames many of those forces as central drivers through 2030, and UN population projections continue to show a world getting older and still growing, but with slower growth than many earlier assumptions implied.

So if you’re reading predictions for 2035, the smart move is to ask different questions than “Is this cool?” or even “Is this possible?”

Ask: Who gets it first? Who gets left out? What infrastructure does it depend on? What rules govern it? What skills does it require? What happens if adoption is uneven for ten years?

That is where the real future usually shows up — not in the headline promise, but in the rollout.

And that may be the clearest lesson from the road to 2025: the future often arrives on time in concept, but unevenly in practice.

What we know / What’s unclear

What we know: Older forecasts often did identify major trajectories correctly (embedded connectivity, AI integration), especially at a high level.

What we know: Global connectivity expanded sharply, but the digital divide remains large in 2025.

What’s unclear: Which current 2035 forecasts will prove directionally right but badly timed.

What’s unclear: How quickly institutions (schools, employers, regulators, governments) adapt compared with the underlying technology.

Sources: Pew Research Center, ITU, World Economic Forum, United Nations.