AI use cases were scattered throughout the news companies featured on the recent INMA Media Subscriptions Summit study tour — a tool to help with the CMS  and text-to-voice here, creating a quiz and assisting the paywall there.

But are these ideas — and others news companies around the world are experimenting with and investing in — transformational enough to satisfy one AI expert on the tour?

AI uses cases among Amsterdam media

Many stops along the two-day study tour, with 27 attendees from news companies around the world, highlighted current or aspirational uses of AI:

At Het Financieele Dagblad (FD), Jeroen Hoorn, head of online, shared that his team made an AI tool to use next to its CMS. The two integrate quite well. The goal: AI-assisted personalisation. 

The company’s radio station, BNR.nl, has an AI-generated news quiz based on news articles from the platform. The just-released product already had thousands of users.

“We didn’t want any human intervention,” explained Julian Verbeek, head of digital. “We wanted a standalone product. We tested it vigorously.”

For example: “How many people died in Gaza last night” wasn’t the kind of questions they were going for. 

At NRC, every news article published can be read to a user with an AI voice.

Mediahuis is planning to add AI and modeling to its dynamic paywall.

At Roularta Media, a team defines AI use cases they want to work on quarterly, Rajno van Wilgenberg, product leader of the AI task force, said. Currently, those are:

1. Content creation: automating text generation, text-to-speech, video creation.

2. Content distribution: adapting interfaces so consumers can interact in a more fluid way.

3. Customer lifecycle management: looking at analytics, creation of analytical dashboards for end consumer to analyse data.

4. Advertising: creating contextual ads.

Roularta is also looking at AI at a corporate level for compliance, content protection, partnerships (monetising bot access and licensing content LLMs), and content discoverability. 

The news publisher is taking a five-step approach to AI adoption: experiments, evaluation, production, implementation, core to ecosystems.

The sooner-than-you-think future of AI

Wesley ter Haar founded Media.Monks 24 years ago and is now also its chief AI officer.

He sees 2025 as a year of transformation for AI. 

“AI changed the economies of many things, but it starts with advertising,” he told study tour attendees. “Our commercial model is mostly selling the hours of people. People and hours equal value. Everybody in this room has a connection to digital advertising.”

He shared a McKinsey report that chose 2,250 jobs that make up the labour of the world, predicting the percentage that would be replaced by AI in the next five years: 30% to 100%, depending on the job. 

The key question is this, he said: “Is this real and is it really happening now?”

Yes. And yes.

Ter Haar mentioned something called Humanity’s Last Exam. As of mid-March, OpenAI can compete with 26% of that exam. The first month, OpenAI passed 9%. A month later, it’s up to 26%.

“After this exam, if machines can pass it, we will never have to make another because machines will be smarter than us. We are two years away from machines being able to operate at PhD levels.”

Mark Zuckerberg has said AI is as good as mid-level software engineers at Meta.

“It’s real,” ter Haar said. “It’s really happening. Muck quicker than it’s almost possible to follow. This is the first time I’ve seen a technology where the promise isn’t just ahead of where most people realise — it’s way ahead of most organisation’s ability to adopt it.”

While everybody eventually will be using AI, Ter Haar said, only a small percentage of companies will figure it out in a way that brings on the full organisation change necessary.

“Many won’t. And that gap will be almost impossible to close.”

He shared what he sees as the three pillars of AI transformation:

1. Talent: giving staff the tools, training, and push to work with AI.

2. Applied AI: automising the internal workflow — “nobody would build your current company [like it is] if you would start tomorrow.”

3. New jobs/renew jobs/replace jobs: new jobs given to AI agents that previously were impossible or not economically via to do; renewing jobs by training AI agents to get as close to the best employees you have; replacing jobs … because that inevitably will happen.

“Does anybody have an existential freakout so far?” he asked.

His advice?

“Internally, think very big and smart small. Educate and celebrate change. For your clients, de-silo, centralise budgets around AI, and cluster per discipline.”

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