The hype cycle is eating itself

The term "artificial intelligence" has traveled a long way from its origins in academic research to become one of the most misused words in marketing. Understanding what happened in between requires looking at how the label got stretched, why companies attach it to everything, and what that means for anyone trying to separate genuine capability from promotional claims. How the label got born The phrase "artificial intelligence" was formally coined in 1956 at a summer workshop at Dartmouth College, where computer scientists gathered to explore whether machines could simulate human intelligence. At that time, the field was genuinely exploratory and the terminology was precise within academic circles. Researchers were studying symbolic logic, problem solving, and early forms of automated reasoning. The concept was specific and bounded: it referred to programmed systems that could execute logic chains and solve defined problems. Over decades, the definition stayed relatively narrow among researchers. AI meant systems that could learn from data, recognize patterns, or make decisions based on explicitly programmed rules or statistical patterns learned from training. The field had clear boundaries, measurable benchmarks, and shared understanding of what counted as progress. When marketing discovered the word By the 2010s and accelerating through the 2020s, "AI" became a commercial label rather than a technical descriptor. Any system that made automated decisions started claiming the title. A software program that sorted emails into folders based on simple rules got called AI. An app that played pre-recorded responses based on keywords was AI. A filter that adjusted photos was AI. The marketing incentive was obvious: the word carried associations with intelligence, sophistication, and technological advancement. It made products sound more capable and futuristic than they might actually be. This wasn't entirely dishonest in many cases. A sorting algorithm technically does make automated decisions. But the gap between what the term meant to researchers and what it meant to consumers had grown enormous. Calling something AI no longer told you whether it learned from data, required human input to function, or had any genuine adaptive capability. Why the inflation accelerated The boom in machine learning and neural networks in the 2010s provided real ammunition for these claims. Systems like large language models, image generation tools, and recommendation engines demonstrated genuinely novel capabilities that people hadn't seen before. These tools could do things that seemed to require human judgment or creativity. The excitement was legitimate. That legitimacy became cover for everything else. If a machine learning model could generate images from text descriptions, then marketing departments reasoned that any product with any algorithmic component could ride that same wave of credibility. The term became so diluted th