In a January 2026 Substack essay, Wired founding executive editor and technology writer Kevin Kelly argues that artificial intelligence is uniquely burdened by what he calls being "Over-Expected" — a concept he builds on Arthur C. Clarke's 1963 taxonomy of anticipated versus unanticipated technologies. Kelly's central observation is a striking asymmetry: the internet, arguably the most transformative technology of the past century, was almost entirely absent from science fiction's imagination before it arrived and was embraced with little cultural resistance. AI, by contrast, has been anticipated, dramatized, and overwhelmingly catastrophized for over a hundred years, with Hollywood producing hundreds of films in which intelligent machines spell disaster for humanity. The result is a public imagination pre-loaded with dystopian outcomes before anyone has meaningful firsthand experience with the technology.
That pre-loading has real policy consequences, Kelly argues. Regulatory efforts aimed at constraining AI before it is fully operational reflect harm as a default expectation, not a conclusion drawn from evidence. He acknowledges these efforts are well-intentioned but argues they are likely futile, since the actual harms — and benefits — of AI remain empirically unknown while the imagined harms feel vivid and certain. The internet contrast is pointed: a technology that arrived without narrative baggage was adopted enthusiastically, while one rehearsed in fiction for a century faces preemptive wariness. Kelly cites Waymo's self-driving vehicles as a counterexample where AI earns genuine public enthusiasm — specifically because riders encounter the technology directly and judge it on firsthand experience rather than cultural priors.
The essay has drawn substantive pushback on Hacker News. One commenter argued that the "shock of obsolescence" many feel encountering capable language models is a rational, experience-based response rather than a fiction-conditioned reflex, and pointed to significant plot holes in <a href="/news/2026-03-14-longitudinal-study-ai-tools-boost-developer-productivity-10-percent-not-hyped-2-3x">optimistic AI scenarios</a> around economic displacement in a post-labor economy. Another rejected Kelly's framing outright, citing <a href="/news/2026-03-14-amazon-employees-say-ai-is-just-increasing-workload-study-confirms">personal observations of AI overreliance harming peers and younger users</a> — pessimism rooted in lived reality, not cinema. A third complicated Kelly's internet-prediction thesis by pointing to E. M. Forster's 1909 story "The Machine Stops" as an early precursor to networked dependency, alongside 1970s science fiction depicting tablet-like personal data devices.
The Waymo insight has direct relevance to autonomous agents: the ones most likely to overcome a century of dystopian priors are those that deliver hands-on value people can judge for themselves. Kelly's closing call — to spend the next decade imagining what AI might get right — is directed as much at technologists and journalists as at fiction writers.