Artificial intelligence (AI) had a starring role throughout the five days of study tours, conference presentations, and in-depth topical seminars of INMA’s World Congress of News Media in New York last week.

One takeaway, said at the end of the New York media study tour, by INMA Researcher-in-Residence Greg Piechota, may be keeping attendees up at night:

“If we lower our costs, have better workflows using AI, if we’re faster at drafting stories, composing videos but everybody has the same tools? Imagine instead of 115 journalists at Newsweek you can have with AI have 2,000 journalists. Or 20,000 journalists. A better question: But what if everybody has 20,000 journalists with AI? 

“What will be the differentiator in a world where every publication has these AI tools and tons of journalists in the field produce tons of content?”

And, as an important aside, he asked: “Do people really want more content?”

AI is a tool, not a replacement for important journalism

Piechota cited findings from a 2024 study showing news organisations are moving toward adopting AI and placing greater emphasis on actively listening to their audiences.

News organisations are now adopting what Piechota’s colleagues predicted years ago when they realised the value of content would decline with digitisation: using AI to produce low-demand stories while reserving the best journalism for readers willing to pay for a subscription.

In taking questions from the INMA audience last week, legendary investigative journalist Bob Woodward acknowledged the growing role of AI in journalism but warned it could also derail reporters if they lean too much on speed over substance:

“If you want to take the patience factor out of journalism and operate on this high speed of what's the latest, sometimes I think you wind up misleading your readers or your viewers.”

At USA Today, AI is not treated as a standalone technology project but as a catalyst for reimagining the entire business model of journalism. The organisation leverages AI to scale its core strength — quality journalism — by automating time-consuming but essential tasks. This frees journalists to focus more on impactful reporting while simultaneously expanding the company’s first-party data through every user interaction.

One practical example is an AI-assisted summary tool that generates article summaries, saving journalists time and improving content reach. Rather than chasing a few large breakthroughs, USA Today views AI as a collection of many small, integrated improvements. With over 100 approved use cases, AI is embedded across functions — from editorial support to data management — driving operational efficiency and enabling new revenue models such as content-as-a-service.

To guide this transformation, the company has established an internal AI council of 80 cross-functional employees. This team is responsible for scaling successful pilots, monitoring regulatory changes, and ensuring ethical, responsible growth. Their priorities include accelerating AI impact, upskilling staff, and sustaining innovation while safeguarding journalistic integrity.

News companies are in an exploratory phase

Almar Latour, CEO at Dow Jones, said he remains in an exploratory phase with AI, aiming to better understand how the technology is evolving and how both employees and consumers might be using it. His team recently tested the tool by posing highly specific questions related to a marketing study, and the results were comparable to those produced by a consulting firm. 

At Dow Jones, the AI team continues to focus on automation and generative AI, bringing together experts in linguistics, computer science, and other disciplines.

“We now scan customer feedback in new ways that will also lead to just setting up customer service and other parts of the business very differently than before,” Latour said. “I think it’s so incredibly important to understand what our customers are saying.”

The Financial Times has adopted a measured and transparent approach to AI in journalism, guided by principles set out by Editor Roula Khalaf in 2023. Since then, the newsroom has created a dedicated AI panel, with cross-departmental representation from editorial, product, and tech teams. This panel oversees AI submissions, approving tools that are logged into a transparent, internal registry.

Several AI applications are now integrated into daily operations, Head of Digital Platforms/Columnist Matthew Garrahan said. These include the AskFT generative search tool for subscribers and an AI Playground where all staff can experiment with prompts and refine outputs. Comment moderation has been streamlined using Ytopia, an AI tool that’s reduced the burden on editorial teams. Meanwhile, AI News Wire clusters related wire entries, helping beat reporters and editors form stronger content habits.

FT has also piloted and gradually deployed automated news summaries, finding they boost reader engagement — particularly when summaries precede full articles. Though still visible only to staff, these summaries are now fully automated, without human intervention, and free of hallucinations. Editors remain involved in ranking and editing where needed.

Beyond newsroom tools, the FT has launched a Computational Journalism team using AI to investigate and analyse complex data sets. Their reporting has uncovered significant stories — from federal contract anomalies to cryptocurrency scandals — that would have been difficult to surface manually. FT’s own generative AI tool, FT Professional, has been rolled out to professional users with 17,0000 questions asked in the last five weeks. 

Despite this progress, FT leadership remains cautious, especially as agentic AI looms. Their focus is on using AI as a reporting tool, not a replacement for human journalism — and on reinforcing the FT’s unique value by deepening its relationship with readers through original, data-driven reporting.

AI uses cases from World Congress

Other AI use cases were peppered throughout the week of World Congress, from study tour stops to the main stage to seminars. Here is a sampling:

• For The New York Times, AI’s value lies in its ability to enhance, not eclipse, journalism. Tusar Barik, senior vice president/marketing, introduced BrandMatch, an AI-powered tool that aligns advertising with relevant content through contextual analysis. This ensures that ad placements are both impactful and ethically sound. At the same time, AI tools support content diversification, including video, games, and other immersive formats. “We first meet audiences where they are,” Barik explained. “We survived by evolving with the times.”

• Gannett has a center with AI journalists who are dedicated to producing and editing massive amounts of content by using AI to turn inputs like press releases and real estate updates relevant to local audiences into stories for everyday readers.

• At the Dow Jones and The Wall Street Journal study tour stop, Tom Gebauer, senior vice president/design at Dow Jones, spoke about the way designers can help consumers understand how new, cutting-edge AI features work — and why they should trust Dow Jones’ AI capabilities with handling their news. “Designers can create experiences that provide users with control and understanding over AI features and how algorithms work, promoting brand and content trust,” Gebauer said.

• The Atlantic has taken a team-based approach to AI integration, creating a collaborative space through its Labs initiative. “We don’t want to be left behind. We can’t put our heads in the sand,” said Jim Quindlen, vice president/product design. Their approach fosters innovation through cross-functional teams of editors, designers, and technologists testing new technologies with minimal risk.

• At Business Insider, journalists have seen a 40% increase in original reporting thanks to AI-assisted workflows that help with everything from court document analysis to interview prep and headline brainstorming. Yet, Julie Zeveloff, newsroom chief-of-staff, made clear, “Our editors and journalists are in charge of everything that gets published, and that isn’t going to change.”

• Yahoo News is using AI to personalise how users consume content. “We now have the possibility to adapt content into users’ preferred format,” said Editor-in-Chief Rosa Heyman. This level of personalisation is not just a convenience but a core engagement strategy. Tanya Pai, director of editorial, added: “We try to differentiate what we’re doing by format … . but we always want to make sure we have that human element.”

• At Dotdash Meredith, Chief Innovation Officer Dr. Jon Roberts warned that while AI can rapidly surface new facts, “only publishers can give nuanced expertise and creativity while ChatGPT cannot.” 

• Condé Nast’s product and technology teams plan to use AI to increase engagement and dynamically surface their most important products, as well as to integrate AI into their internal workflows: “We need to know our audiences and have data about them to serve them better,” Patrick Gray, executive director/commerce, said. 

• Using AI, News24 can match advertiser content with relevant editorial topics, increasing the chances of discovery and engagement. “We’ve found that when advertisers want to highlight something new or time-sensitive, contextual placements work incredibly well,” said Jerusha Roth, publisher of News24. But context isn’t the only strategy. Roth explained how even “Know”-oriented advertising can serve more complex goals, like launching a new product. In those cases, News24 offers partners a bespoke experience that mirrors the newsroom environment.

• Scientific American is still struggling to bring data to the newsroom, Kimberly Lau, media general manager and business development executive, told study tour attendees. “How do we take advantage of GenAI and use it to our advantage? We’re a little bit behind on this. Our editorial team is very anti-AI in many ways, and some of their concerns are warranted.” The newsroom does use AI for transcriptions and translations.

• At Newsday, “content is leading the way, but our goal is to have AI leading us in everything we do,” Chief Product Officer Bharat Krish said. Generative AI searches have increased 1500% from 30,000 to 500,000 per month. AI also provides 75,000 translations per week at the news company. “What is the latest with the U.S. tariffs? AI can answer that.” One day, Newsweek’s entire homepage could be based on an answer engine, he said.

• Reuters, which has an AI oversight committee led by its editor-in-chief, has an AI tool called LAMB (Lynx Automated Media Packager), which cuts down production time by 8%, and another called LUX Summary, which creates a headline, overview, and other article features. These ideas come from problems to be solved by staff members. “One of the most satisfying things in this AI rollout is just discovering all these colleagues who I would never in a million years think would be leaders in AI,” Mahesh Ramachandran, head of technology, said.

• James Martin, executive director of new products and innovations at The Independent, outlined how the publisher created “Bulletin,” a brand for time-poor readers. It uses AI-generated bullet point summaries and key headlines to give users digestible news updates — all while preserving editorial integrity. “We spend hours, and days, and months drafting fantastic original journalism. But if we can’t get people to read that long-form journalism, wouldn't we want them to at least read a version of that journalism?”

• To guide 7,000 employees across six countries, Mediahuis launched an internal programme called AIM (AI + Mediahuis) to increase the adoption rate of AI with newsletters, a Slack channel, videos explaining concepts like “agents,” and frequent hackathons.

• Nikkei has developed Ask Nikkei, a chatbot trained solely on its own content. If no relevant article exists, the bot simply doesn’t answer — prioritizing accuracy and trust. It can answer within five seconds and provide clearly sourced explanations. Pre-generated questions help the system maintain editorial control and performance. 

Thoughts from news company leaders on AI

Stefan Ottlitz, CEO at Spiegel Group: “AI is going to be everywhere in our products.”

Julie Zeveloff, newsroom chief of staff at Business Insider: “We have an open-minded newsroom that’s excited to find new ways to incorporate AI into our work.” 

Andrew Morse, president and publisher at The Atlanta Journal-Constitution: “You have to understand which of your products and which of your capabilities are going to meet different audience needs. AI can be incredibly effective in helping you reach those different audience segments and turning them into paying customers.”

Michael Reed, CEO of Gannett: “If you’re running a business and you're not starting to think about how to apply AI to every facet of the business … you're making a huge mistake.”

Tony Hunter, McClatchy CEO: AI is part of a broader effort to speed up operations and reclaim time for storytelling at McClatchy. Quality content and audience differentiation are top priorities. “Everything else, I’m willing to automate, outsource, and run at the lowest cost possible so that I can continue to invest in those two areas.” 

Sally Buzbee, news editor/USA and Canada at Reuters: “I think many newsrooms have been so scared of AI and of the technology that is disrupting our industry in so many ways. I am just really blessed to be at a place that is positively engaging with how we can make this new technology help our journalism and amplify our journalism. There’s a million opportunities in terms of surfacing content, curating content, and making archives more efficient.”

Gerold Riedmann, editor-in-chief, Der Standard: AI should be deployed strategically, not just for the sake of it: “In some boardrooms, I have heard the new KPI is how much of your content is AI-touched.”

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .