How will AI reshape the news in 2026? Forecasts by 17 experts from around the world

As we enter 2026, and the third year since the transformative release of ChatGPT, journalists and media managers are wondering what the next frontier for generative AI and the news will be. We got in touch with some of the most prominent voices working in this space (and put out an open call to our audience) to get a sense of what this year might bring.

An obvious and important caveat: neither our respondents nor we have a crystal ball, and nobody knows for sure what the future holds. Nonetheless, we found five recurring themes in their forecasts:

• Audiences will increasingly access news through AI
• There will be increased demand for verification work
• Automation and agents will reshape newsrooms
• Newsrooms will upskill and build AI infrastructure
• AI will further empower data journalists

There were additional themes that came up in the audience responses, but not in our expert submissions. These include the idea that AI will take on a role in commissioning and setting the news agenda, changes in business models to cope with a collapse in direct traffic to news websites, and a shift in trust and authority from brands to individuals. 

Below, we expand on these five themes, with a summary of the views of our audience and the thoughts of each of our 17 experts in full. 

Jump to forecast from: Gina Chua from Semafor | Alessandro Alviani from SZ | Olle Zachrison from BBC | Ezra Eeman from NPO | Sannuta Raghu from Scroll | Florent Daudens from Mizal.ai | Joshua Ogawa from Nikkei | Shuwei Fang | David Caswell | Katharina Schell from APA | Tess Jeffers from WSJ | Tshepo Tshabalala from JournalismAI | Sebastián Auyanet Torres | Rubina Fillion from NYT | Sonali Verma from INMA | Martin Stabe from FT | Jaemark Tordecilla

1. Audiences will increasingly access news through AI 

As audiences increasingly use generative AI-powered chatbots, search or other AI-informed tools when searching for information, traditional ways of accessing the news will give way to news stories being discovered and accessed via AI tools. This was our most prevalent forecast overall: a plurality of both our expert forecasts and our audience contributions fell under this broad category.

1.1. From our audience:

“The broader picture will be declining traffic, but discerning, high-intent readers will increasingly use LLM apps not for full stories, but to decide which to consume and why. Core news audiences will choose from more diverse outlets on the authority or access of individual articles, not brands,” wrote Nicola Leech, head of audience development at Emirati newspaper The National.

“For news organisations, this shift is bigger than AI writing stories,” wrote Mweha Msemo, Tanzania correspondent for Finnish magazine Maailman Kuvalehti. “They lose control over how journalism is presented. There is no front page or fixed order. AI breaks articles into pieces and uses only what it needs.”

Many responses mentioned an “answer economy”, which would replace article-first consumption and allow audiences to personalise news by directly asking AI models how something would affect their lives. 

“People will consume news less by reading articles and more by asking an AI assistant and receiving it in their own context, like 'Explain the impact in my life,’ ‘Summarise this for my industry,’ ‘How solid is this claim?’ Audience behaviour will split into two modes: comfort mode, summary and suggested actions; and trust mode, a demand to see evidence, sources, and quotations. For both, journalism will become a layer inside the ‘Answer Economy’,” wrote Cigdem Oztabak, a journalist for CNN Turkiye.

1.2. From our experts:

Gina Chua
Executive Editor at Large, Semafor

Audiences will accelerate their use of chatbots/LLMs to access information, despite their well-documented issues with accuracy and hallucination, and traffic to news sites will continue to fall. 

Some newsrooms will attempt to migrate upstream, banking on brand reputation, star journalists and the newsroom's voice to build direct relationships with loyal readers. Others will turn to AI for efficiencies to increase their output. And a few enterprising newsrooms will begin experimenting with how to tap into this new user behaviour by providing chatbot-like interfaces to users.

Alessandro Alviani
Lead for Generative AI, Süddeutsche Zeitung Digitale Medien

After decades of optimising for visual discovery, 2026 could mark a turning point: screenless, audio-based conversational experiences may increasingly become an entry point for information consumption. The lines between 'reading,' 'listening' and 'interacting' will blur. Users will be able not only to ask questions and receive quick answers but also to switch to linear audio for deep, original reporting seamlessly. 

This creates new opportunities but also challenges. The first is structural: how do we design journalism for situations where users move smoothly between conversation and long-form listening without using a screen? This goes beyond the often-invoked ‘death’ of the article format. It is about building new user flows and experiences that do not yet exist, at a time when a new AI-mediated information ecosystem has profoundly redefined traditional habits and entry points.

Olle Zachrison
Senior News Editor AI, BBC News

In 2026, we’ll hear even more about AI-powered browsers and device-level AI modes. Features like Google’s AI mode, ChatGPT’s Atlas mode and Microsoft’s Copilot sidebar will gain more traction and increasingly reshape how audiences consume news. These tools bypass efforts to block AI crawlers – users can simply ask their device to explain, summarise or translate whatever is on their screen. 

This further speeds up the decline of search referrals, but it also has implications for media companies’ own AI-powered services. If summaries, translations or conversational features become built-in features of devices – like new generation earphones – that must be factored into our own design choices. 

The news industry will likely need to focus on more bespoke AI-assisted solutions, both internally and for audiences, rather than mimicking commercial assistants. This is intrinsically positive, as it helps safeguard independence, uphold editorial standards and retain better control over our data.

Ezra Eeman
Strategy and Innovation Director, NPO

AI models will become a sticky layer across everything in 2026, absorbing time and attention, not for news in the narrow sense, but as catch-all interfaces for information and entertainment. Publishers will realise it’s not about adding AI to their workflows, but getting themselves added to AI. Moving from "AI in Media" to "Media in AI." 

As big tech integrates AI across entire ecosystems, it will become an inevitable gateway to media. People will no longer need to choose between searching and scrolling: AI will blend both, surfacing what you need before you know you need it. Content will match your moment, mood, and context. 

For publishers, the options are limited: embrace AI assistants and figure out how citations are part of a new economic model, or provide exclusivity and relevance beyond the summary. Close to communities and deep in expertise, these news brands will offer what giants can't easily replicate.

Sannuta Raghu
Leader of Scroll Media’s AI Lab

The most significant shift will be the collapse of the idea that one article equals one story. The article has always been a closed container, a single linear object, because print and URLs demanded it. It assumes uniform users arriving at the same moment with the same level of knowledge. In reality, audiences arrive with very different familiarity, questions, and needs.

As malleable interfaces are built into news products, the article becomes an entry point. Each article container can pull in contextually relevant material from across a newsroom’s entire archive based on readers’ needs. It can also expand to include additional features such as podcasts, short videos, etc.

For audiences, the change is experiential: people won’t “read the news” so much as navigate and query verified information. For news organisations, the shift is less about changing how journalism is done and more about changing how information is organised for use.

Florent Daudens
CEO, Mizal.ai

In 2026, chatbots will become the new app stores. We're already seeing it: OpenAI announced ChatGPT will surface third-party apps directly in conversations, and the Model Context Protocol is becoming a standard to plug any service into AI assistants.

For news organisations, this is a distribution shift hiding in plain sight. Just as publishers had to figure out Facebook and Google, they'll now need to be discoverable inside conversations. Clicks won't be the measure of success. Conversation will be, which means figuring out new value chains and user experiences.

This might prove tricky, but there’s an opportunity gap: A recent Reuters Institute study showed information-seeking has become AI's primary use case (24% weekly), yet news consumption sits at a mere 6%. Part of the equation lies in connecting people with individual voices, not just institutional identities.

2. There will be increased demand for verification

Credibility will differentiate news outlets in a world where information is easy to access, but trust is low. Audiences will want to see evidence and sources to back up what they are told online, and news publishers have an opportunity to meet this need. This idea was more popular among our audience contributions than among our expert forecasts. 

2.1. From our audience:

“‘Breaking verification’ will replace ‘breaking news’ in 2026, and trust will decide who survives,” wrote Vinay Sarawagi, co-founder and CEO of The Media GCC, a media operations outsourcing partner company based in India. “A digital chain of custody could become a costly signal of truthful information,” wrote Jannes Jegminat, a postdoc in clinical machine learning at the Icahn School of Medicine at Mount Sinai in the US. A digital chain of custody is the trail of documentation that details each step in the transfer, control and analysis of a piece of electronic evidence. 

2.2. From our experts:

Joshua Ogawa
Head of AI & Visual Strategy, Editorial Division, Nikkei

Whether you like it or not, seeing is no longer believing in the age of AI-generated slops and deepfakes, which are flooding the digital information space. Photo and video journalism are no exception.

News organisations generally had a very low tolerance for any digital alteration of images or videos they publish, even before the advent of generative AI. Now that anyone can easily and cheaply create photorealistic synthetic images/videos, leaving little evidence in the process, it is increasingly difficult to keep the same journalistic standard.

Expect the news industry to finally get serious about investing in the tools and skills required to verify and authenticate visual content. The adoption of available solutions such as C2PA has been painfully slow for various reasons. But we still have time to preserve visual journalism’s role as proof and avoid the liar’s dividend.

Shuwei Fang
Shorenstein Fellow, Harvard Kennedy School

This is the year in which news organisations wake up to their next product: not content, but process. An opportunity arises for somebody to create a product that responds to the question 'Is this real?' at speed and with credibility.

Why now? Synthetic content has already flooded the information environment. 2026 is when it turns adversarial. In August 2025, nearly half the social media outrage over US restaurant chain Cracker Barrel's logo change was synthetic; authentic criticism amplified into a stock-tanking controversy. Expect this to mature and become intentional: micro-targeted, orchestrated attacks designed to move markets and extract value.

Audiences won't learn to spot fakes (they can't), so they'll delegate, and some are willing to pay [for this service]. AI-powered detection is already sold by defence contractors, but to a narrow market. The news media have the editorial credibility to take this further, but may lack the agility. Startups have the agility, but must build the trust.

3. Automation and agents will reshape newsrooms

This theme encompasses deeper and more comprehensive integration of AI by newsrooms: AI will become embedded in newsrooms’ CMS and workflows. Publishing of some information will be automated and carried out by agents, and stories will be automatically updated. The much-discussed ‘human in the loop’ might be quietly retired.

Not everyone views this as a positive development, with some in our audience, particularly students and young people, voicing concerns about job cuts and rushed AI adoption.

3.1. From our audience:

“Use of AI by organisations to churn out the news in the fastest way possible,” wrote Saumya Singh, a master’s student in human rights at the London School of Economics and a former sub-editor and producer. “By 2026, the biggest shift will be AI running entire workflows, shaping the news experience for audiences, and forcing all of us, especially those just entering the field, to confront what integrity looks like when the ground won’t stop moving,” wrote young journalist Pablo Urdiales Antelo.

3.2. From our experts:

David Caswell
Consultant

2026 will see news organisations increasingly use agentic AI for the end-to-end automation of complex workflows. 

In 2023-24, many news producers automated individual newsroom tasks like summarising articles, generating headlines, drafting newsletters, copy editing and the like – often deployed via ‘AI toolkits’. This produced some useful efficiencies. But by 2025, the limits of ‘task automation’ have become apparent. Savings of time and money are underwhelming, and task-focused AI seemed like a strategic dead-end.

Meanwhile, AI agents enabled by new ‘reasoning models’ have appeared – processes that understand broad goals, ask clarifying questions and then execute the many individual tasks needed to achieve those goals. ‘Deep Research’ tools are early examples.

AI agents can already automate very sophisticated knowledge production workflows that far exceed the complexity of simple newsroom tasks. In 2026, more newsrooms will discover these powerful capabilities and begin using them strategically in newsgathering, investigations, interviewing, fact-checking and more.

4. Newsrooms will upskill and build new AI infrastructures

2026 could be the year in which newsrooms invest in the infrastructure and training necessary to make the most of what AI already has to offer. This is particularly relevant for small newsrooms, which may not have dedicated roles or investment yet.

4.1. From our audience:

“The AI problem for newsrooms isn’t about tools. It’s about systems. The current media system was built around journalism as a craft, not as a product that can scale and adapt technically. 2026 can be the year of infrastructure when systems finally start catching up. You can’t just drop a new engine into an old machine and expect it to run,” wrote Sergei Yakupov, founder of AIforNewsroom.in, an index for AI uses in newsrooms, and Novocean.me, a media consultancy specialising in AI implementation. 

4.2. From our experts:

Katharina Schell 
Deputy editor-in-chief, Austria Presse Agentur

News organisations will shift their focus from AI in production to AI in distribution and monetisation. The potential of AI in media production has been partly overestimated and partly already exploited, but content monetisation becomes an increasingly acute challenge. Even media companies that have been reluctant up to now will consider content deals with AI platforms in 2026. 

Conversational AI will continue to grow when it comes to up-to-date information. ‘News will find me’ will gradually be replaced by ‘I can request news at any time’. As a paradoxical side effect, the public will continue to lose confidence in news disseminated online, as the high degree of AI penetration means that virtually all information is suspect. 

Tess Jeffers
Director of newsroom data and AI, Wall Street Journal

2026 will be the year news publishers fully leverage generative AI to better serve their audiences. 

Publishers will use synthetic audience models – AI chatbots trained to embody key audience personas. These “always on” sounding boards will deliver reporters and editors instant feedback. Do you want to chat over an idea? Check if your lede hooks your target reader? Just ask your (AI) audiences. 

News personalisation will hit its stride, moving beyond content to generative AI-powered customisation of format, tone, style and depth. While audiences rarely want every feature, most readers will use one or more of these options at some point. News publishers will build a flurry of GenAI products to give audiences this flexibility. 

The final step is the full democratisation of audience data across the newsroom, powered by dedicated data chatbots. Data and insights will no longer be confined to dashboards or specialised roles. Instead, audience intelligence will be instantly available to everyone. 

Tshepo Tshabalala
Project manager, JournalismAI

AI will stop being just a fun experiment and will become a basic necessity for smaller newsrooms.

In 2026, these small and medium-sized news operations will be relying on AI primarily to become more sustainable and save time. Think of AI being a super-efficient digital intern: it’ll handle the boring, repetitive tasks like summarising long articles, transcribing interviews, and crunching simple data. This frees up the human reporters to focus on the serious, impactful stories for their respective communities. Crucially, AI will also start helping them become more sustainable and think about implementing revenue growth strategies to stay afloat.

The big catch, though, isn't the technology itself, but getting people to use it right. Small newsrooms will still struggle to make AI work perfectly for very local stories and to get their staff fully comfortable. The smart players will invest in training and ethics, using AI to do better work. 

Sebastián Auyanet Torres
Consultant in audience development, product and impact

The most significant development in 2026 should be the mainstreaming of "vibe coding" – using natural language to build customised internal tools. Newsrooms of all sizes  – and individual news creators as well – will increasingly move from generic software to architecting their own inbound systems perfectly suited to their operational reality. This capability might unlock, for example, superior first-party data collection and precise audience identification.

However, the ultimate impact is human. By automating complex logic through custom tools, we eliminate the operational excuses for isolation. This technology finally frees us to do what algorithms cannot: get out of the building to listen, feel, and facilitate connection. In 2026, AI becomes the engine that powers a return to face-to-face community service that has already been happening in different ways.

Rubina Fillion
Associate editorial director of AI Initiatives, New York Times

While at the New York Times, we never use AI to write articles, it can help create first drafts of summaries and metadata. Even these brief snippets of text must meet high editorial standards. When newsrooms use AI for SEO headlines or alt text, they need concrete ways to measure if they’re actually hitting the mark. 

Our AI Initiatives team works with journalists and product leaders to develop frameworks to evaluate editorial quality. That starts with deciding on key characteristics, like accuracy, and how to score them for each bit of text generated by AI. That provides us with data to improve prompts and select the right models for the task. The copy is thoroughly edited before publication, in accordance with our principles for using AI. 

Good writing is subjective. But there are still ways to measure it. My colleague Duy Nguyen and I shared what we learned from this process.

Sonali Verma
GenAI Initiative Lead, INMA

Using AI for efficiencies in news will become old hat: The real focus will be on revenue generation through new products that help us serve existing audiences better and bring new audiences in the door. We will see news organisations imagining products that were until now, incredibly expensive or impossible to create (e.g. multimodal and personalised) and monetising them effectively.

5. AI will further empower data journalists

Data journalists have long utilised AI, and as the technology improves and becomes more accessible, it could enable newsrooms to gather and sort through an unprecedented number of documents. This was mentioned twice in our expert forecasts, but was only a minor theme in our crowdsourced responses. 

5.1. From our experts:

Martin Stabe
Data Editor, The Financial Times

One great promise of AI for reporters is that it will enable them to trawl documents at scale. But to find a needle, you first need to assemble a haystack. And proactively collecting troves of potentially newsworthy data for analysis is not something that most news organisations have historically done.

Sure, data journalism teams have long ingested vast datasets for individual stories. For these tech-savvy teams, investigations based on LLM-enabled document classification have already become routine output.

The rest of the organisation typically maintains just one major dataset for editorial use: their own archives. So it’s no surprise that so many media AI applications focus on summarising and otherwise repackaging previously-published stories.

But almost by definition, archives are not where scoops are to be found. For that, you need fresh data from external sources. More newsrooms will realise this in 2026 and embrace new editorial-facing data engineering functions.

Jaemark Tordecilla
Journalist, technologist, and media advisor

Processing and publishing public datasets that are AI-ready will be incredibly valuable for both news organisations and news audiences, given how easy it is to investigate them with the help of chatbots. 

Earlier this year, when the Philippines was in a furore over flood infrastructure corruption, I vibe-coded a script to scrape data from the government’s official website and posted it publicly as a spreadsheet. It allowed media outlets and think tanks to create their own investigations. More interestingly, regular citizens took that spreadsheet, uploaded it into ChatGPT, and asked questions about it in their local contexts.

With thanks to everyone who shared their thoughts with us.