In 2026, I can no longer assume that you are experiencing my prediction through these words that I’m writing.

Instead, you may grasp the gist of my point through an AI-generated summary, or listen to two fictitious people discuss its implications as a podcast, or even watch a video of two cartoon cats re-enacting it as interpretive dance. (I’d actually pay good money to see this.)

Generative AI has proven adept at transforming content into different formats, lengths, and tones of voice. These transformations will become more prevalent in 2026 and will require journalists to shift their mindset about how information is designed and delivered.

This change is akin to the move towards responsive web design when smartphones gained popularity. Instead of being able to dictate the placement of every pixel on a fixed-sized screen, designers had to think more about modularity, break points, and the underlying structure of their designs. It initially felt restrictive relative to what they could create purely for the wider desktop canvas, and they had to learn to be comfortable with widowed or orphaned words being out of their control.

Before we herald a new era of “Responsive Content Design,” there are limits to this analogy. Responsive web design was fundamentally about layout and how information is presented; generative AI enables the transformation of what that information is, its meaning, and its accuracy.

The transformations that generative AI makes to information are not lossless. Content that becomes malleable loses some fidelity to the original. This loss can be through a divergence from authorial intent, from the accuracy or truthfulness of specific claims (i.e. hallucinations), or by straying from the voice or style of the original.

Some of this ‘loss’ may be intentional (“tl;dr, just give me a summary”). Some may even make the information more accessible, informative, or engaging (for example, this interactive adventure version of a history book, or translating to different languages). But the mere fact that content is now easily malleable, and can be molded to fit each individual’s preferences, will only introduce more uncertainty about the reliability of journalistic content, and further erode trust in news.

I can speculate about a future where factchecking-as-a-service is ubiquitously and seamlessly integrated into such transformations to ensure that factual claims are preserved no matter which format news is transformed into. Or a more dystopian view of fragmented reality and even stronger filter bubbles as a result of this new technology. But on the more immediate horizon, I see two main responses that journalists, as publishers of information, can have to this shift:

• Lean into the malleable nature of content and its new possibilities. Use AI to generate the summaries, the podcast versions, the short-form videos in a controlled and useful way for users. At the same time, create content specifically designed to be molded by machines: AI Search Content optimization, Tyler Cowen on writing for LLMs.
• Create, label, and celebrate experiences that resist malleability. Walter Benjamin, in The Work of Art in the Age of Mechanical Reproduction and later John Berger, in Ways of Seeing, both wrote of the aura and uniqueness of an original work of art that distinguishes it from its reproductions. Journalists have traditionally tried to ensure that news can reach as widely as possible, but there may be value in conveying truths through unique and limited experiences. For example, returning to print or live journalism performances.

In 2026 and beyond, both will have their place in the output of forward-thinking publishers. In the meantime, if someone has good tips for writing that interpretive dance prompt…