The past is gone: AI under the microscope as the latest Innovation in Media World Report is launched at the FIPP Congress
The overhaul of digital discovery, the rise of direct reader relationships and AI’s power to unlock creative and collaborative horizons are the main themes of the Innovation in Media 2025-26 World Report, launched at the FIPP World Media Congress in Madrid this week.
Authored by Juan Señor and Jayant Sriram of Innovation Media Consulting, the annual report – published in partnership with FIPP – delves into the most profitable developments impacting media today and is compiled using a vast network of consultants around the word. Print copies were handed out to delegates at Congress while the digital version is available to order here.
Covering a wide range aspects of the media landscape, the report looks at publishers as a philanthropic partner, as an educator, a brand licensor, a club, a purveyor of games, a think tank, a betting agency, an affiliate marketer, an archivist, an IT provider, an event organiser, a marketplace and a vanity publisher.
Addressing the issue of our times, Señor warned delegates they were facing an existential crisis because of AI.
“I see a bit of denial in the industry as to what’s coming, and it worries us. I think it’s important to talk about it openly and put it on the table,” he said. “The past is gone and we’re at this moment in history – I call it a chiaroscuro moment.
“We really don’t know what the future holds, but we can confidently share with you that the past is gone. And we work so hard to build a model and we finally have a model in digital and I’m afraid it’s gone. The internet is dying in front of our eyes, in slow motion.
And our businesses are built on that internet. Our websites, apps, browsing, search – it’s dying. And we need to really think deeply about what’s to come.
“As we move firmly into this AI age, everything is at stake, it’s existential. Madrid 2025, for all of you, it’s going to be an existential moment. Because the future 100 years of publishing really belong to those who defend the worth of our words.”
Building bridges to audiences
This year’s report explores what is a growing anxiety amongst publishers when it comes to declining traffic to news sites, fewer social media referrals, the broad rollout of AI features in search engine and ongoing uncertainty about the programmatic advertising model.
According to Señor, the best way to address these concerns is for publishers to prioritise and build direct relationships with their audiences.
This direct connection would, ideally, lead readers to subscribe and create a strong revenue proposition. However, the transition is complicated, and concerns have been mounting for years that the current subscription model is plateauing.
But, despite the slowing growth in subscription revenue, there are still opportunities to return to growth by employing a mix of offensive and defensive strategies, highlighted in the Innovation Report.
“ChatGPT has got the attention of the world. They’ve got the attention of our readers,” said Señor.
“And let’s also agree that it’s being designed to replace visits to our sites, to our assets. It’s a traffic killer. So how do we reinvent the business to assume zero clicks – a Google zero strategy in terms of revenue from search? Well, that’s something that should be discussed and addressed immediately.
“Platform traffic is gone, so assume a business model without that traffic. And I think it’s a prudent way to plan your business going forward. Because I think that’s something that we can confidently say – the direction of travel is irreversible.”
How to make money with AI platforms
This Innovation report also delves into the high-profile negotiations happening between media companies and Big Tech – pointing out that publishers are in a complicated position when it comes to partnering with AI platforms.
On one side, licensing deals with companies like OpenAI and Perplexity offer a way to generate new revenue, ensure attribution, and even access advanced AI tools. On the other side, the benefits aren’t always as clear-cut – or as great – as they first appear.
The situation is made even more complex by the fact that, in both the US and EU, regulators are now examining whether AI companies should be required to obtain explicit consent and pay for content they use to train models.
Proposals under discussion could standardise licensing frameworks, strengthen attribution requirements, and make enforcement against unauthorised scraping easier. For publishers, this could tilt the balance from “nice to have” to “non-negotiable”.
Señor reminded delegates that their quality content gave them leverage with tech companies.
“We are 30% of what goes into the mix, right? And they rank us – OpenAI just announced that we are AAA. They’re willing to pay. They need trustworthy content. And we provide it. And it’s wonderful to hear them say this publicly. They acknowledge that we’re worth a lot.”
Señor stressed the importance of publishers making their content AI readable. “In the generative AI age, it’s about being the answer. And this is a different craft. The past was built on links, we must be built on answers.
“And the revenue should be diversified from the subscriptions we have, all that reader revenue, to licensing platforms, but also new formulas like pay-per-scrape – when the LLMs find a request for something that you have, there is now a method for them to pay you just to access that.”
Going beyond programmatic
The Innovation report further looks at the shifting advertising landscape – looking at the decline of traditional ad channels, discussing alternative strategies, and considering what the future holds for publishers amid these evolving challenges.
For over a decade, digital advertising was dominated by programmatic buying and search-driven traffic. These channels once provided a steady, reliable revenue stream, but the landscape has shifted dramatically.
With the gradual phase-out of third-party cookies – initially expected by 2024, but now extended – along with evolving privacy regulations and the rise of ad blockers, traditional ad channels no longer deliver the same results. At the same time, algorithm changes and the emergence of AI-powered search tools have further eroded the traffic that once drove advertising revenues.
“How do we go beyond programmatic scale? We’ve been chasing that scale for so long, and intimacy with audiences is the new currency,” said Señor.
“You can sell much more with a verified, intimate audience than you could ever with those CPMs at that volume – that is just falling. So, this is tomorrow’s goldmine, right? We must get that first-party data. And that requires registration.”
AI in the newsrooms
The Innovation report devotes a large section to one of the biggest talking points in media circles – AI entering the newsroom.
The reports shows that newsrooms can’t just introduce AI tools and expect success; they need to invest in teaching staff how to use them, updating workflows, and setting clear guidelines.
Despite a number of initiatives – including Mediahuis setting up an AI task force with a central coordinator who trains and supports the adoption of AI tools – AI adoption remains low in many organisations, with some reporting that only 25-30% of journalists are actively using AI tools.
A key takeaway to come out of the Innovation report is not to focus too much on short-term profits. Some executives warned that obsessing over quick returns can kill innovation before AI’s full potential is realised.
Instead, they suggested starting with small, practical uses – like using AI to handle routine tasks and free up journalists’ time – before expanding to bigger, revenue-driving applications.
Señor described AI in the newsroom as “the intern who never sleeps, but who needs constant supervision.”
“AI’s doing wonderful things. AI is an incredible set of tools. It does so many things, but it need supervision. And we’ve seen OpenAI admitting that OpenAI will forever lie, will forever hallucinate.
“There’s a percentage that’s less than 1% that will always hallucinate. That’s just the way the algorithm works because it has to search for answer. It always has to give you an answer. So sometimes it has to combine things that perhaps do not compute. So that constant supervision is key throughout.”
