Journalism is in danger of bending itself around AI

There’s a popular myth that journalism has missed out on the last waves of technological change because it was too slow or too ignorant. The fear of disruption still looms large in an industry increasingly dependent on and battered by a handful of tech companies. Reality, of course, is more complex and messy. Media companies had websites before anyone had even thought of Google; they built audiences on social networks before “influencer” was even muttered for the first time.

The industry’s “misses” were often structural, economic, or cultural rather than technological: resource limitations, pressures of short-term profitability, reliance on existing business models, and the difficulty of scaling quality journalism sustainably all played a role. And the fact that tech companies were completely unburdened by all these.

But media companies have adapted. As a recent paper by James Meese and Theresa Seipp observes, media companies are increasingly following the logic of big tech while trying to maintain independence. Over the past decade, outlets have become obsessed with data strategies, personalization, and scale. Or to put it another way: the mistake was corrected, though its messiness, not simple novelty, dictated that process.

Still, news companies are seemingly hell-bent on avoiding the mythical mistake with “AI”: adopt, adopt, adopt.

But the question I find myself asking is whether the media risks overcorrecting: distorting itself to fit the affordances of a new technology rather than letting these tools serve journalistic values. This is a recurring pattern with a certain form of hyped technologies: they act like black holes, bending priorities and behaviors around themselves. Instead of thinking about how a new tool might fit the profession, the profession instead reimagined itself around it. We have seen this in the past with calls for things like “journalism on the blockchain” or “journalism in VR.”

But there’s a consequence to this reimagining. Once newsrooms are seen as primarily a set of distinct tasks, “AI” can simply be slotted into, they become managed like a set of pipelines. Journalism, in this view, is suddenly treated as a kind of liquid, something to be portioned, mixed, shaped, and moved through the system to maximize flow (or “productivity”) and reach, rather than being carefully curated for context or meaning. Journalism becomes simply content. Something that somehow flows from one end to the end-user, while “AI” is placing the valves, even though humans might still turn some.

In one meeting I recently attended, a manager claimed that “the purpose of a newspaper is to deliver high-quality content in the form readers prefer.” While not entirely wrong, this misses journalism’s democratic role: informing the public, scrutinizing power, and sustaining a functioning society. But it perfectly encapsulates this view through the pipeline, where delivery and form supersede democratic function

This may be the inevitable outcome of media companies being forced to adapt to the market logics dictated by tech companies. Still, we should be careful with what the end product looks like — especially considering how these market logics are currently inflating yet another economic bubble. The danger is that by liquifying, journalism risks becoming endlessly malleable, shaped to fit yet another set of companies’ platforms, metrics, and algorithms rather than public needs. The question isn’t whether journalism should use “AI” tooling, but whether it can do so without losing sight of why journalism matters in the first place and what distinguishes it from mere content production. Is it worth the risk?