Insights
If your audience is reading your story through Chat GPT, journalism has a problem.
August 2026
Originally written for fathm as part of their series of thought pieces from their Expert Network
During a conversation with a young undergrad at a London university summer school, she mentioned offhandedly that she drops everything she wants to read into Chat GPT.
That’s what she does after finding something she thinks she might want to read. Drops it into a third-party tool to ‘decipher’ it for her.
At first, this made me feel incredibly sad. What about the delicately crafted language? The nuance? The journalist’s intention?
Then I felt scared. Did Chat GPT give her an accurate reflection of what the article intended? Was she being told things which weren’t true?
Then bewildered. Why would you even do that? What’s the point? Did she have too much of a popcorn brain to read it for herself?
After the rush and weightlessness of this rollercoaster of emotions I finished up frustrated and angry. At myself. This isn’t a ‘her’ problem. This is an ‘us’ problem – the ‘us’ being the journalism industry.
This intelligent, bright young person found more value in the Chat GPT re-evaluation of the piece than she did of the publisher’s. She expects to get nothing from their interpretation of the story, so she finds her own, one that works for her. The Chat output gives her what she needs – whether it is brevity, context, a format which helps her understand. Whatever it is, she gets what she needs from the transaction and goes on her merry way.
I only spoke to her for a few minutes, and it was humbling. I already know from work I’ve been involved with that journalism is struggling to meet the needs of our audiences, but hearing it put so plainly was still hard to hear.
I’d met her at a discussion along the lines of – what has AI in journalism done for us? I’d spoken about the opportunities in the workflows of reporters and newsdesks; about translation, recommendations, finding order among disorder for investigations. Why this all helps audiences get better content. But better for whom
During the discussion, I’d shown the group this chart, which had surprised me when I’d seen it in the Reuters Institute Digital Media Report 2026.
This young person had shown me how true this can be.
When I asked her why, she simply said, it just helps me understand it better.
And quicker.
It was like a dagger to the heart. Her experience is that news stories don’t give her a full understanding of the context of the story. AND they take too long going about it. If I was in her shoes, and my needs in that moment weren’t being met, I would filter out the noise too.
The Reuters report goes further into why people choose to go through Chat GPT, Claude etc to re-interpret the news for them.
For journalists, the results make pretty bleak feedback.
Summarisation: stories are too long;
Follow up questions: we don’t offer a good enough onward journey or explain the context of the story well enough;
Make stories easier to understand: who are we writing for if people need to put it through another tool to make it understandable?
Reliability of the news source: The Report, quite rightly, points to concerns over press freedom in some countries. But interestingly, it raises another aspect – it’s not the disengaged who use chatbots to determine the reality behind the story.
The Report suggests the use of chatbots in this way is concentrated among those already engaged with, and have a high interest in, the news – those who want to be able to interrogate it further.
It’s highly engaged people with agency and literacy who are applying these extra layers to aid their understanding.
I’ll need to bring another aspect of AI into the conversation here. It can be the code to unlock the door. I see AI as a tool in the journalist’s toolbox, next to the screwdrivers, the natural curiosity, and the ability to ask – “but why?”
As a key, it can be used to reformat, rework, redeliver our content to different audiences in the formats they find most palatable, to help them understand what’s happening and in what context. It’s something we looked into at BBC News Labs before AI made it easy. With the usual caveats, I would love to have a homepage which truly reflects me and my interests – different kinds of content, stories from around the world, in different languages, in formats I find simple to synthesise. I can get this from an aggregator right now, but if publishers want audiences to go to – and stay – on their platforms, then they’ll need to give them what they are looking for.
At the start of these discussions with undergraduates, I often canter through why I think I have the privilege of being able to stand at the front of the room.
Twenty-five years of experience. Newspapers. Radio. 24-hour rolling news. Digital. Innovation. I hope I have enough self-awareness to know that I can be a conduit, maybe even a catalyst in conversations to help people think more, create more, build more.
But everyone needs a 19-year-old undergrad on their shoulder – I now have mine – to remind them, who are we really doing this for?