Who Decides What Counts as an AI Revolution?

An antique radio beside a book and lamp, illustrating how an older medium becomes part of present-day reflection.
Photo: Esra Erdem via Pexels

A technology can change the way we work without changing everything people say it will change. I keep returning to that distinction whenever a new communication tool is described as a revolution.

The 1938 War of the Worlds broadcast offers an older example. Some listeners were frightened by fictional news bulletins, while newspaper coverage made the incident into a much bigger public story about radio's influence. The episode mattered for what occurred and for what it supposedly demonstrated about audiences.

A claim about transformation needs evidence of what changed, not only confidence that change is coming.

When an event becomes a lesson

The history becomes more interesting when we look at the institutions producing research about it. Pooley and Socolow (2013) examined archival evidence about the influential research that followed the broadcast. Their account shows how disputes over scholarly credit and interpretation helped shape the episode's place in communication research.

This does not erase the listeners' fear. It reminds me that the story of what a technology does can also have a history, including choices about which findings receive attention and how they are remembered.

Today, similarly broad language circulates about generative AI: that it will transform writing, eliminate familiar tasks or redefine expertise. These are different claims, and each requires its own evidence.

The Rockefeller Archive Center's records also show that the broadcast attracted attention from research funders and educators worried about propaganda and public judgement. The episode did not speak for itself. Organisations selected what questions to investigate, and their interests shaped which problems seemed especially urgent.

A task is not a profession

One useful comparison comes from Noy and Zhang's (2023) experiment with college-educated professionals completing specified writing tasks. Participants given access to ChatGPT completed the tasks faster, and their work received higher quality ratings in that setting.

That result is meaningful. It supports a claim about performance on particular tasks, not a forecast about the disappearance of journalists, teachers or other professions. Those changes depend on organisations, resources, responsibilities and decisions the experiment was not designed to measure.

A faster first draft could reduce workload, change an editor's expectations or simply create more material to review. Which of those possibilities occurs cannot be settled by timing one writing task.

For writing work, this leaves questions about who decides whether speed is an improvement. If an assignment values quick summaries, a faster draft may be welcome. If it values independent reasoning, attribution or careful interviewing, the same measure says less. Different institutions can define a successful result quite differently.

Who gives change its meaning?

What connects the radio controversy and present-day AI discussion is the difficulty of moving from an observed event to a confident account of its wider consequences.

I explored the historical comparison in a this EMPOWERVERSE article. What interests me here is how easily a measurable result can become shorthand for an imagined future.

These claims also leave their sources and become conversations about work, learning and expectations.

When the next warning about AI seems persuasive, perhaps the first useful question is which part has been observed: its output, someone's response, or a story about what everyone else must be thinking.

Popular posts from this blog

Carl Jung in 2026: The Persona, the Shadow, and the Search for Wholeness

The Architecture of Alienation: A Psychological Dive into Society's Fragmentation of Shared Reality

The Dangerous Allure of Pop Psychology: A Critical Analysis