Why is AI important?
AI matters because it makes cognitive work cheap for the first time — reading, drafting, translating and analyzing now cost close to nothing per unit. Half of American adults already use chatbots (49%, up from 33% in 2024). It's also reshaping physical infrastructure: data centre electricity use is projected to more than double to about 945 TWh by 2030.
Why — the first-principles explanation
Strip away the hype and one economic fact remains: AI collapses the cost of a certain kind of thinking. For all of history, anything requiring reading, judging, summarizing or writing required a human hour, and human hours are expensive, finite and don't scale. That constraint quietly shaped everything. Contracts go unread. Medical images go unreviewed. Small businesses have no legal department. Not because the work is impossible but because the hours cost too much.
When a cost falls by a factor of a thousand, you don't get "the same activity, cheaper." You get activities that were never worth doing at all. Nobody photographed their lunch when film cost a dollar a frame; when the cost hit zero, we invented Instagram. So the important question isn't "will AI do my current job faster." It's what happens when it becomes economical to read every contract, check every scan, translate every document. Most of that hasn't happened yet — the technology arrived years before the institutions figured out what to do with it.
The second-order effects are where importance actually lives. If AI is cheap, it gets used everywhere, and everywhere includes decisions about your loan, your job application, your medical triage. Then the interesting questions aren't technical — they're about who's accountable when a system nobody fully understands makes a call about your life. That's why the EU wrote a law with fines up to €35 million or 7% of global turnover, and why 71% of Americans think AI will make personal information less secure.
And it's important physically, which surprises people. Thinking has always been the one thing that didn't need a power plant. Now it does. The IEA puts data centres at 415 TWh in 2024 — roughly 1.5% of world electricity — growing about 12% a year since 2017, more than four times faster than overall electricity demand, and projected to more than double to about 945 TWh by 2030. AI is the most important driver. That's a real constraint being negotiated in real grid planning meetings.
Honest caveat: "important" is not "good." Pew finds 40% of Americans predict a negative societal impact over 20 years and 63% say AI is advancing too quickly. Something can matter enormously and still be a mess.
An example that makes it click
Think about what happened when writing became cheap. Before the printing press, a book cost a monk a year. Books existed, but only for people who could afford a year of a monk. After Gutenberg, the price collapsed — and the result wasn't just "more Bibles." It was newspapers, pamphlets, novels, scientific journals, literacy, and eventually a couple of revolutions. Nobody in 1450 predicted any of that from a cheaper Bible.
AI does the same thing to reading and judging rather than copying. That's why the honest answer to "why is AI important" is partly: we don't know yet, the same way 1450 didn't know.
Key facts
- 49% of U.S. adults use AI chatbots as of February 2026, up from 33% in 2024 and more than double the 2023 share — Pew Research Center, 5,119 adults, February 17–23, 2026.
- 24% of U.S. adults use chatbots daily; 38% of employed adults use them for work tasks.
- Data centres consumed about 415 TWh in 2024 — around 1.5% of world electricity — growing about 12% per year since 2017, more than four times faster than total electricity consumption (IEA).
- The IEA projects data centre consumption will more than double to roughly 945 TWh by 2030, slightly more than Japan's total annual electricity use, with AI the most important driver; scenarios span 700–1,700 TWh by 2035.
- 40% of Americans predict AI will have a negative societal impact over the next 20 years; 63% say AI is advancing too quickly; 71% believe AI will make personal information less secure (Pew, 2026).
- The EU AI Act, in force since August 1, 2024, sets fines up to €35 million or 7% of worldwide annual turnover for prohibited AI practices — a measure of how consequential regulators judge the technology to be.
▶ The 60-second explainer (script)
AI is important for one economic reason: it makes a certain kind of thinking almost free. For all of history, anything that needed reading, judging, summarizing or writing needed a human hour. Human hours are expensive and finite. That constraint quietly shaped the world. Contracts go unread. Scans go unreviewed. Small businesses have no legal department — not because the work is impossible, but because the hours cost too much. Here's the thing about costs collapsing: you don't get the same activity, cheaper. You get activities that were never worth doing. Nobody photographed their lunch when film was a dollar a frame. Cost hit zero, we invented Instagram. So the real question isn't whether AI does your current job faster. It's what happens when it becomes economical to read every contract and check every scan. Most of that hasn't happened yet. The second-order effects are where importance really lives. Cheap means everywhere, and everywhere includes decisions about your loan, your job application, your medical triage. Then it stops being a technical question and becomes: who's accountable when a system nobody fully understands makes a call about your life? That's why Europe wrote a law with fines up to 7% of global turnover. And it's important physically. Thinking never needed a power plant before. The IEA puts data centres at 415 terawatt-hours in 2024 — about 1.5% of world electricity — growing 12% a year and projected to roughly double to 945 by 2030. But important isn't the same as good. Pew found 40% of Americans expect a negative impact over 20 years, and 63% say it's moving too fast. Something can matter enormously and still be a mess.
What authoritative sources say
People also ask
Is AI important, or is it a bubble?
Both claims can be true at once. Adoption is real — half of U.S. adults use chatbots, and data centre buildout is showing up in national electricity forecasts. Whether current company valuations match that is a separate question, and one that history says is usually answered badly at the time.
Will AI take my job?
The credible estimates disagree sharply, and the honest answer is that nobody knows. What's better established is that AI changes tasks within jobs faster than it eliminates whole jobs — Pew finds 38% of employed U.S. adults already use chatbots for work tasks. Beware anyone quoting a precise percentage of jobs lost; those numbers come from models with very different assumptions.
Why does AI use so much electricity?
Because it runs on chips in data centres that must be powered and cooled continuously. The IEA puts all data centres at about 1.5% of world electricity in 2024, projecting a doubling by 2030 with AI as the main driver. Note that figure covers all data centre activity, not AI alone — AI's exact share isn't cleanly separable, which is why estimates vary so much.
What's the strongest argument that AI isn't that important?
That most of the value hasn't materialized in measured productivity yet, and that chatbots being popular isn't the same as chatbots being economically transformative. Adoption curves and impact curves aren't the same curve, and the gap between them has embarrassed a lot of past technology forecasts.
Do most people think AI is good?
No. Pew's 2026 survey finds 40% of Americans expect a negative societal impact over 20 years, 63% say AI is advancing too quickly, and 71% think it will make personal information less secure — even as usage climbs.