Every generation has its miracle technology.
It arrives amid breathless predictions, billion-dollar investments and promises that it will change our lives. Skeptics are dismissed as people who simply “don’t get it.” Then reality sets in. The technology isn’t a failure, but it isn’t the revolution we were promised.
Artificial intelligence is simply the latest example—although its economic consequences could be far greater than anything that has come before.
For the past several years we’ve been told AI would soon rival, even surpass, human intelligence. Some predicted it would replace programmers, lawyers, teachers, artists and even doctors. Others claimed it would eliminate nearly all jobs. Investors responded by pouring hundreds of billions of dollars into AI companies, chip manufacturing, and the massive data centers needed to power them. Businesses, afraid of being left behind, rushed to bolt AI onto virtually every product they sell.
Yet anyone who uses AI regularly knows the reality is far more complicated.
AI can be remarkably useful but it isn’t nearly as intelligent as advertised. It still hallucinates, invents facts, misquotes sources and occasionally produces answers that are simply wrong. Customer service chatbots still misunderstand basic questions. AI assistants still require careful supervision before anyone can trust important decisions to them.
None of that means AI lacks value. Quite the opposite.
For me, AI has become a useful research tool. Instead of searching dozens of websites, I can ask for a table comparing every electric vehicle under a certain width that will fit in my garage, and receive an organized table in seconds. It summarizes lengthy reports, organizes scattered information, and often explains complicated subjects better than a conventional search engine. (I used AI to create the image for this column.)
Those features make AI a genuine breakthrough. But it isn’t artificial intelligence in the sense we’ve been promised. It’s more like artificial assistance.
We’ve seen this pattern before.
When compact discs arrived in the 1980s, consumers were told digital audio was the future. CDs were more durable, easier to use and allowed instant access to any track. Vinyl records, we were assured, would soon disappear.
Instead, vinyl made a remarkable comeback. Listeners rediscovered qualities they preferred, while CDs proved to be more of a tradeoff with its own audio issues. CDs were more convenient, but not better.
Google Glass promised wearable computing that would replace smartphones. It never happened.
Self-driving cars were expected to make human drivers largely obsolete years ago. Instead, fully autonomous vehicles remain limited to carefully controlled environments.
The pattern is the same with all of these. Many technology companies today no longer simply introduce products. They sell visions of the future. Those visions attract investment, boost stock prices and convince businesses they can’t afford to be left behind. The bigger the promise, the easier it becomes to raise money.
AI is following the same script. The difference is its scale. Never before has so much money been committed so quickly to a technology whose practical limits are still becoming clear. Hundreds of billions of dollars are flowing into AI infrastructure, particularly giant data centers, based on expectations that AI will transform nearly every aspect of the economy. But what if its greatest strength turns out not to be replacing human intelligence, but simply making people more productive?
History suggests that’s the more likely outcome.
