Proof of how Gen-AI is transforming the news ecosystem is evident in the many case studies featured at WAN-IFRA’s World News Media Congress in Krakow.
Like: how it’s engineering a shift in audience behaviour, says Nikita Roy, Harvard-recognised AI futurist and founder of Newsroom Robots Lab, who delivered a presentation that revealed both thrilling and alarming developments.
She illustrated this with a simple ‘before and after’ graphic that clarifies how AI is transforming the search behaviour, collapsing it from multi‑click journeys to single conversational answers, replacing the familiar blue‑link workflow.
Roy had analysed how newsrooms are using chatbots to engage audiences, across the world, over the past year. “I started seeing patterns of jobs that each of those chatbots are starting to do,” she added, emphasising the customisation of tools that pushes personalisation to a higher level, by providing contextual info for specific user needs.”
How newsrooms are using Chatbots to help audiences
UNDERSTAND News easily: Launched just this week, Sweden’s Aftonbladet helps people find answers from its news archives in 50 languages. “What makes it interesting is that we can now talk to it, in two languages,” explains Roy.
EMPOWER Spotlight PA’s Election Assistant enables your audience to make informed decisions.
NAVIGATE events: The Hindu, India launched Margazhi in December, to helps users find content schedules across multiple different venues and events.
DISCOVER: San Francisco Chronicle’s Chowbot is built upon food and wine coverage.
On the surface, these are basically content recommendations. “But instead of being able to scroll to something, now you’re able to get contextual information for your specific need,” explained Roy.
Waves of change: From text to voice
This is just the first interface emerging with Gen-AI: “The next big shift that I’m seeing coming is voice-first, conversational,” she added. “This is not just chatbots: Your audience is talking.”
Gen Z and Gen Alpha are steering this growth, with their embrace of AI.
“An eMarketer study found that 64% of the US Gen Z population will be using a voice assistant monthly by 2027 – up from 51% in 2023,” noted Roy.
‘Amazon Alexa answers 25 million questions from children every month and Character.AI, which has a predominantly Gen Z audience, receives 20 000 queries per second – a fifth of Google’s volume.’
“This audience behaviour from Gen Z and Gen Alpha is showing that no longer is information a broadcast – it’s now a dialogue; a conversation that is showing intent and ability to find the information that meets the intent and need at that moment,” Roy advised.
Capitalising on a trend
It’s also setting a trend, with venture capital funds pouring money into voice agents: “In December ChatGPT was one of the first AI tools to go multimodal; in the US, Google is converting that ‘Discover’ seat into a five-minute podcast. And Microsoft Copilot took this one step further just last month (April) by generating interactive podcasts on demand, on any topic.”
Roy illustrated the power — and disruption — of new AI interfaces using the example of Microsoft Copilot’s podcast capabilities. She described how the tool can automatically create interactive podcasts on any topic, pulling in information from across the internet, including her own work on Newsroom Robots.
It will create a podcast on a chosen topic based on content and information provided. “And I can take that down a separate route and have a conversation with my podcast hosts about it, about the topic,” she explained.
Growing audience demand
Roy emphasised that this kind of generative experience fills a growing audience demand — but it also raises important questions for newsrooms.
“It’s meeting a personalized need using content I’ve created,” she said, “which is why I built an AI version of myself — the Nikita AI clone — using my public work, podcast episodes, trainings, and talks. She’s available 24/7, in more than 30 languages, and delivers both chat and voice-first experiences to guide audiences.”
This, reckons Roy, “is the brand coming alive; Newsroom Robots coming alive to help guide audiences to find that information for them.”
Now for the synthetic audience
Prepare for the end of the article, warned Roy. “The next big paradigm shift we’re seeing are AI agents becoming your audience, and when the computer decides to take actions on your behalf.”
Roy pointed to ChatGPT’s “Deep Research” feature as proof of what agentic AI can already do. Instead of forcing users to click through a maze of blue links, the tool scans news sites it has partnerships with, gathers the relevant material, and drops it all into one tidy output.
Things get really interesting with Chat GPT’s Operator, which can mimic human behavior by navigating news websites and comparing content across homepages. She also highlighted how other tools — like Convergence AI — go further, using prompts and workarounds to summarize paywalled content via services like Archive.ph.
“This is no longer about scraping websites; we are entering another conversation that we ought to be having,” advises Roy, illustrating how Agentic AI can instruct the model on how to bypass the paywalled article.
“It’s taking human-like actions, and I’ve seen this happening consistently across almost every newsroom I’ve tested. I haven’t yet found a newsroom where I couldn’t bypass the paywall using this method.”
It sees data points
All of this is pointing to the death of the article as we know it, adds Roy: ”Because when an AI model, or agent looks at an article, it’s not seeing content; it’s seeing data points and, in trying to make sense of a bunch of different data points, it’s remixing everything and creating a completely new set of information.”
What next for an industry that has been slow to understand and respond to the many different transformations required by this digital era – and in danger of losing our future audiences?
Roy set this challenge: “I hope that we can begin to really tap in across our organisations – not just to the journalists, but to our product teams, audiences, data teams – look into your own newsrooms, build things that you can measure, build things that help you, and build your generational audiences.”
This story was first published by the World Editors Forum, a leading global community of editors, which is an integral part of the World Association of News Publishers (WAN-IFRA).
Lucinda Jordaan is WAN-IFRA’s correspondent in Africa.
Further Reading
- Shibsted Future Report 2024: On speaking terms with machines
- More about Synthetic Audiences and Personas for product development and testing
- This AI expert is creating a chatbot to keep you safe
