Artificial Intelligence has overhauled the possibilities around personalisation. During the recent INMA Generative AI Master Class, Sonali Verma, lead of the INMA Generative AI Initiative, looked at the many ways that is playing out in news media companies around the world.
“What we’re now seeing is nothing short of a revolution,” Verma said. “Everyone is experimenting with personalisation, and it doesn’t cost an arm and a leg anymore.”
This has led to what is being called “an era of liquid content,” in which GenAI can effortlessly change the language and format of the content. And that is handing the reins of control to the audience.
“What’s significant is that it is now the news consumer who decides what their news experience will be, not the editor,” Verma said.
During the master class, experts from Time Magazine, JP/Politikens Group, The Telegraph, Scroll.in, and Google Labs shared how AI is being used to create dynamic content experiences for audiences.
Time enables readers to chat with its Person of the Year
Time Magazine’s unveiling of its Person of the Year has become a much anticipated and globally watched event: “It is one of, if not the single most important and most anticipated, editorial franchise announcements that we make each year,” Mark Howard, chief operating officer, said.
That presented a unique opportunity to take “our biggest announcement that we know is going to have the most global eyeballs and do something truly transformative.”
Time introduced a multifunction AI toolbar that rethinks the user experience and allows them to interact differently with the content. And the annual Person of the Year cover provided the ideal opportunity to showcase it.
Time made it possible for readers to chat with its Person of the Year via its AI toolbar.
Partnering with the San Francisco-based company Scale AI, Time focused on creating a new experience for the December 2024 unveiling of its Person of the Year, which was then president-elect Donald Trump. In addition to Trump, it deployed its AI Toolbar for the previous three years’ recipients: Taylor Swift (2023), Volodymyr Zelensky (2022), and Elon Musk (2021).
However, things became more interesting when Time implemented voice chat, which allowed users to hit the Speak With Me tab and ask questions about the personalities featured in the content.
Howard said that although the chat offered just seven languages, people would communicate with it in their native tongue, and the technology was sophisticated enough to respond to them. In all, 24 languages were used, which drove higher engagement than was seen amongst people who simply visited the site for the text.
Based on the success of this initiative, Time will start training a private LLM on its entire 102-year-old archives and plans for a “bigger, broader release of this product” later this year, Howard said: “You’ll start to see it on our main features, our cover stories, our exclusive features.”
JP/Politikens Media Group balances technology, editorial integrity
For JP/Politikens Media Group in Denmark, using AI for personalisation is a delicate balancing act Kasper Lindskow, head of AI, said. If they think the way they’re using the technology to give readers what they want will compromise their values as an organisation, they’re not going to do it that way.
In 2019, some in the newsroom started realising AI technology was to a point where they probably should start dabbling. Today their mission is to help shape the way news is produced and consumed by enabling departments to responsibly use AI in ways that will advance the individual strategies of their different news brands.
They’re structured in a way where a centralised unit serves all the brands across the group. That unit is made up of technical specialists and product managers that have backgrounds in both journalism and tech. This helps them bridge the gap between AI solutions and the editorial departments.
“And then we collaborate with AI hubs in the editorial rooms on each brand to try to make sure that we can adapt stuff to the needs of the individual brands, but also that we have some ambassadors out there that help us roll out the systems so that they’re actually used,” Lindskow said.
Brands under the JP/Politikens Media Group umbrella are able to tailor experiences for their specific audiences by using AI.
AI allows them to get more granular with what works for individual brands and how to balance their needs. When they tried personalisation in print newspapers, they were really only able to personalise with their mix of news content in each edition.
Then, for the first 15-20 years of digital publication, they had more of a one-size-fits-all approach to personalisation. And now, in an AI era, they can personalise news flow and not just the content itself.
“Now we are also working on doing it more with Gen AI so inside the content units this is a lot of benefits because we are much better able to serve the individual needs of readers and make sure that all the content that we find important to report on is actually read by someone,” Lindskow said.
The Telegraph augments content with AI
At The Telegraph, a U.K. institution with a long and storied history, there is a clear shift towards a more subscriber-first approach. Jason Sheppard, senior product manager for machine learning, emphasised the crucial role of personalisation in transforming news media: “Personalised content ensures that readers see stories, topics, or formats that align with their preferences.”
The Telegraph is leveraging AI across a variety of formats.
One significant area where AI adds value is in the augmentation of traditional articles with data-driven visualisations. Sheppard discusses how AI can enhance articles with relevant data insights, providing deeper context and enriching the reader experience.
“Key data insights, a very easy way to augment data into visualisations that is coming from the article or adjacent topics,” he said.
These visualisations not only make the content more engaging but also help present complex information in an accessible and understandable format.
AI’s capabilities are also harnessed to improve the accessibility of content through narration and translation, transforming written articles into more digestible formats like podcasts or multilingual versions. This approach significantly broadens the reach of their content, making it accessible to non-English speakers and those who prefer auditory learning.
“Especially adapting it to the spoken word, so it feels like the author’s talking directly to you, rather than having a one-to-one translation from text to speech,” Sheppard said, adding that this approach enhances the personal connection between The Telegraph and its readers.
Scroll.in transforms news formats
Sannuta Raghu, head of the AI Lab at Scroll.in in India, said the outlet faces unique challenges that underscores the necessity for innovative approaches to reach a broader audience: “We operate in an ecosystem which is largely controlled today by conglomerate-owned media … .We’re an English-language small newsroom with under 20 journalists.”
AI is helping the company automate and refine the transformation of news formats. This process includes developing algorithms capable of condensing complex articles into simpler, more digestible formats that cater to the needs of a varied audience.
“We are testing currently if you’re able to do this at scale, if you’re able to do this with context awareness, and does it have fidelity to source in every generation,” Raghu said.
Raghu highlighted two particularly innovative explorations: The “Calculator” and the “Context Slider,” each designed to meet specific user needs by transforming the presentation and interaction with information.
Scroll.in’s AI-driven calculator can give readers personalised insights derived directly from content.
The AI-driven “Calculator” is tailored for topics where readers can benefit from personalised insights directly derived from the content. For instance, during financial news coverage like budget announcements, this tool can simplify complex fiscal data into a straightforward, user-friendly interface that calculates individual tax liabilities based on the latest fiscal regulations.
The “Context Slider” allows readers to adjust the depth of information they receive from an article based on their existing knowledge and interest level. This tool dynamically modifies the content presentation, offering everything from a succinct overview to an in-depth analysis, depending on the reader’s preference.
Google Labs facilitates flexible content consumption
Kawandeep Virdee, a creative technologist at Google Labs, said potential advancements in AI interfaces offer a vision of the future where news consumption becomes a more interactive, engaging, and personalised experience.
Virdee showcased a prototype where AI dynamically generates summaries at varying lengths, allowing users to delve into a topic as deeply as they wish. This concept alters the static nature of news articles by offering a customisable, layered reading experience.
A prototype demonstrates an AI-generated content summary that can be tailored to any desired length.
“And it can go smaller and smaller until I get just a two-word summary of it,” Virdee said. “It’s interesting because then you just feel like the text becomes more malleable. You can imagine it fitting into different contexts.”
This flexibility in content interaction introduces a novel way for readers to engage with news, making it adaptable to different reading preferences and contexts.
YOU MUST BE A MEMBER TO ACCESS THIS CONTENT
Become An INMA Member or Log In
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
