Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Will 2025 be the year that users embrace generative AI-enabled news?
There’s no shortage of players betting that it’s coming. Ask Perplexity about some current event, and it’ll helpfully summarize the information from a half-dozen news sites, including links so you can check its accuracy for yourself. MIcrosoft’s Bing and Google’s Gemini — and others — will do much the same thing. And The Washington Post recently launched Ask the Post AI, a chatbot that answers questions based on the Post’s reporting and archives.
Whether readers cotton on to this way of getting news remains to be seen. But if and when they switch from reading news to having stories created for them by AI on the fly, it’ll herald a seismic change in the news industry — upending how we work, what we create, and how we make money. We’d be smart to get ahead of it.
To be sure, there are many hurdles to overcome before a world of GenAI-created news becomes commonplace, not least questions about how accurate such systems are. But it’s also not hard to see the attractions — for users, at least — of such a world: A world, at least in theory, where you don’t have to hunt through multiple stories and sources to get a range of perspectives, including from foreign language-publications; where context you already know doesn’t clog up copy and background you’ve forgotten is just a query away. A world where systems know what you’re interested in and how much time you have to read.
The technology to create much of this world is already largely here; but is this something we should welcome? It’s possible many more people are much better informed in a world where AI agents scour the web and find and summarize information that matters to them, giving them access to far more news than they can currently access; it’s equally possible we’re all sucked deeper and deeper into information silos as we read only what is created for us.
The only thing that’s certain is that the world of journalism, and the work of journalists, will change profoundly. Sure, there will always be a place for deep investigative work, sparkling writing, and carefully crafted, insightful narratives. But that’s not what most journalism is; we’re largely in the much more mundane business of informing readers about that day’s events. And that’s the work that AI systems will disrupt first.
What’s the value of a narrative story if it isn’t widely read; if, instead, its facts, insights and analysis are extracted and reworded by an AI system? Is there any point to being the third news organization to confirm some piece of information? Will news brands blur into each other in AI-summarized stories, even if prominent credit (and links) are given to each organization? How can advertising- and traffic-based business models survive if there’s less and less need to visit news sites? Will readers be happy to stay in siloed subscription gardens if they can get multiple views and perspectives from AI systems?
How should journalists and news organizations adapt to this world? Some things come to mind. True exclusive information — scoops — will be critical. Smart, unique insights will be hard to replicate. Voice and personality can attract and retain readers. It might not hurt to rethink what a “story” really is. In-person events — “Live Journalism,” as we call it at Semafor — can’t, at least for the moment, be carried out by chatbots. And deep engagement with, and knowledge of, the community you serve is one way to keep them loyal. (Luckily, all things Semafor is leaning into.)
But I’d be lying if I said I knew what will work in a brave new world of AI news. I only know that we should be actively preparing for it.
