The journey to AI: Tim Forrest on making ITN’s archive ready for the future

Artificial intelligence is dominating discussion across the media industry, but much of the attention remains fixed on the eye-catching end of the technology. Generative AI systems capable of creating text, images, audio, and increasingly convincing video inevitably grab headlines. For ITN, however, there is another AI story taking place. It is less about generating new pictures than finding, understanding, and making better use of the real ones it already has. 

Tim Forrest, Head, Content Delivery, Commercial Innovation, ITN, is preparing to discuss that journey at his session “AI in Practice: Case Studies Across the Media Value Chain",  looking at the work required to make ITN’s enormous archive ready for the opportunities presented by AI. “A lot of IBC talks is about the bleeding edge of AI and how adoption is being taken to places that we’ve never been before,” he says. “We are looking at how we move really central parts of ITN’s business to a place where it’s better positioned for AI, and that is, in many ways, a harder journey than the small POC or the team that is rightly looking to push boundaries. The story I’m trying to tell is about the journey from nowhere to somewhere, and that is actually a really important journey that ITN is embarking on for its archive.” 

Grounded in reality 

Forrest’s perspective on AI is closely connected to the nature of ITN itself. Its business is founded on the value and credibility of real-world pictures. “One of the things that people immediately reach for when they think of AI is generative AI, and actually one of the important things that ITN has established right from its motto ‘Truth to Life’ is that we are a company that is grounded in reality,” he says.

He adds: “Everything we do in understanding AI needs to be thought about through that prism of not wanting to weaken or dilute the importance that we place on reality. One of my primary jobs at ITN is thinking about how we syndicate or sell content around the world. If people believe my content is not real, it’s not worth what I think it’s worth. So, it’s really important to me in an AI journey that we make that distinction to start with, and that although AI can be incredibly supportive, it doesn’t actually take over our product. It is not our product. It assists our product.” 

The immediate opportunity is around discovery. “I’ve got a 70-year archive. We passed 70 years back in 2025, so we’re coming up on 71, but next year it’ll be 72. Every single day, that archive is growing bit by bit, which means that the picture needle that you’re looking for in our haystack is getting ever harder to find without additional support,” says Forrest. “We know that there are great AI tools that have been around for a very long time and are very mature. Think transcription, think translation, think scene detection. All of these are very mature technologies, and all of them help discovery in a real and engaged way.” 

Until recently, however, ITN faced a fundamental obstacle to taking advantage of those tools. Its internally developed media asset management system had been in place for around a decade. “It’s a good system. It’s based on keyword search. It helps you find a lot of what you want, but does it help you find exactly what you want? Are there now systems that will help you get there better?” he asks. 

“During 2025, we ran a project with a company that brought AI systems and techniques in, and we found that we couldn’t integrate it into our internal system because it was 10 years old and was designed in a world before APIs. We determined that our system was effectively end of life, and that to move into the world that we can now see is reaching all parts of our business in one way or another, we needed to move to a different system, and we needed a system that could help us with those mature AI technologies. But we can see what’s becoming an increasingly reduced timeline as the technology matures quicker and quicker.” 

Beyond search 

Forrest believes the technology is already beginning to move beyond straightforward discovery, “I think we’re already seeing a maturity in technology that helps you with early parts of production,” he says. “Even this time last year, AI was very good at showing me exactly the picture that I wanted. Now it’s beginning to take my script and say, ‘You might like this picture against the next phrase and that picture against the third. That doesn’t mean that it’s taking over. It still needs a human to come in and work with it and to think about ‘well, actually, you may think that’s the best picture, but I don’t.’ It’s like a junior picture researcher’s early work but sped up.” 

ITN is not claiming to have reached that point operationally. “What we are trying to get to is a place where we can consider that and consider whether it’s right for our organisation,” says Forrest. “We’re not there yet, but the journey that we are turning up to IBC to present is that without the underlying technology that enables all of that, you’re just nowhere.” 

From cataloguers to curators 

Automation inevitably raises questions about jobs, particularly when AI is involved, but Forrest rejects the idea that ITN’s archive project is primarily about replacing people. “When you’ve got a 70-year archive, as I have, there is always too much work to do, and so the idea that I am replacing people is for the birds,” he says. “I’m revenue-driven in my job, and we can only see growth, but that growth is going to be through a mixture of leaning into technology where it’s relevant and also repurposing my archivists to do more of what helps that growth in different ways.” 

That means reconsidering the archive itself. “The more we understand about our archive, the more we realise that to date it has sort of been thought about as a single entity, and yet actually, what pulls people into the archive are the many and varied stories that exist across it,” says Forrest. “Although you can put whatever discovery tools you want against it, that’s only as good as the question that someone ends up putting into it. Therefore, trying to encourage people to see what they don’t know is a really human endeavour. Drawing out those stories, we’ve begun thinking about our archive in collections, rather than thinking about it as a single entity.” 

“We’re moving away from the idea that you just turn up to someone and say, ‘What would you like?’ Although that question can still be there and it’s still perfectly relevant, and we certainly need to serve that. But we’re moving to a place where we’re trying to say, ‘have you considered?’ What we need to do is think about curation of our archive in a very different way, and think about how we encourage our archivists not any longer to be just cataloguers and data input people, but actually active curators of the archive, drawing people towards what excites and what inspires, and that really is where I see the people that I employ bringing real value, because ultimately this is an archive for other people, and I employ the archivists because they’re really passionate about all those stories and all the great pictures that exist in it.” 

Forrest sees AI removing some of the work that prevents those archivists from spending more time with the material itself. “I want to empower them to no longer be stuck behind a keyboard having to put in metadata manually. I want them to be able to help me surface those great collections that exist right there through our archive.” 

Context remains human 

ITN has another advantage when evaluating new systems: a substantial internal customer base. Its newsrooms serving ITV, Channel 4 and Channel 5 regularly use the archive and related services. “Because we run three newsrooms, who are always using our archive or our services in one way or another, we can get feedback instantly,” says Forrest. “We’re able to literally go to the coffee machine and ask people what they think and get honest feedback.” That proximity also exposes one of the more difficult problems facing AI in journalism. Referencing an IBC Accelerator project he says, “The whole the whole Story Object Model discussion is about context and about how the tricky thing about getting our tools to understand us is that they don’t understand the context that that wraps around everything.” 

Forrest began his career as a journalist and spent much of it producing programmes. “I came up as a journalist. I’ve spent the majority of my career producing programs. You could tell when there was a big story going on just by the sound when you walk in that room, and you know almost the feeling. It’s something in the air.” 

That human layer remains difficult to capture technically. “I think it is one of our primary challenges around AI integration into news for certain, but probably a lot of live production is how to draw that context in in a meaningful way, so that our tools are as up with us as we are with the story,” he says. 

And for Forrest, there is a more fundamental reason why people remain at the centre of the process. “Journalism at its base level is people telling stories that somehow, in one way or another, impact other people. Why would you try to get rid of people in that loop?” 

The journey to AI 

Forrest’s IBC message is consequently less about a dramatic AI breakthrough than the work necessary to make useful adoption possible. ITN is modernising the infrastructure beneath its archive, improving discovery, looking at how automated tools can assist early production, and shifting the role of its archivists towards curation rather than manual data entry. “I guess, in summary, our story is a simple one,” he says. “We are recognising that we live in a world where AI is a huge part, but we’re not getting carried away with it to the point that we’re letting it take over our product. We are letting it assist us and assist our people.” 

To hear more from Forrest and his fellow panellists, don’t miss the IBC2026 Conference session titled “AI in Practice: Case Studies Across the Media Value Chain" taking place on 13 September, 14:00 - 15:15, Conference Room 1. 

To learn more and register for the IBC2026 Conference, click here.