Jyri Kivimäki leads the AI development and transformation at Yle, the Finnish national public broadcasting company. He will hold a workshop on “AI in journalism and newsrooms” at the 2026 ESMH Summer School on science communication in the age of artificial intelligence. We asked him a few questions about the transformative role of AI in journalism and what participants can expect from his workshop.
Can you tell us a little about yourself and what sparked your interest in artificial intelligence and journalism?
Jyri Kivimäki: I’ve been at Yle – Finland’s public broadcaster – for over 20 years now. I started as a sports journalist in the early 2000s, and from the very beginning I was among the first to push for the company’s digital transformation. I’ve always been fascinated by how the role of a journalist evolves when the world and technology around us change. If I look at the common thread in my career, it’s the constant urge to explore and stretch the boundaries of what a journalist can be and what new skills the job requires. Whether that was mobile journalism, deeper interactive storytelling, or today’s AI possibilities.
I currently work as the Head of AI Development and Transformation for Yle across the entire organisation. When I got the chance to start exploring the impact of generative AI on journalism at Yle in the spring of 2023, I jumped at it. Now, roughly three years in, we have built the foundations for Yle’s AI work, taught AI literacy across the organisation, and co-created entirely new solutions for how journalism is done. And honestly, the pace just keeps accelerating…
Narratives around the impact of AI on society often swing between two extremes: apocalyptic fears or utopian promises. Where do you stand?
Jyri Kivimäki: Every coin has two sides, and the extremes rarely represent the truth. There’s that old saying: ‘people tend to overestimate the impact of technology in the short term and underestimate it in the long term’. I think that applies here. AI is built by humans, and humans are responsible for how it’s used. It’s up to us to decide how we want to participate in this development and what responsible AI use looks like for our organisations and our societies. As a public service media organisation, we don’t necessarily have to be the first to implement every new AI trend, but we do have a responsibility to use technology efficiently to serve our customers as well as possible.
And from your experience, how is AI transforming journalism – both in terms of reporting and newsroom practices?
Jyri Kivimäki: AI is a massive shift for journalism. It makes things possible that simply weren’t before – like analysing huge volumes of data. At Yle, for example, we have used it to help reporters comb through tens of thousands of policy and legislative documents and identify newsworthy items in the haystack. Journalists still make all the decisions and heavy lifting; AI supports this by making it possible to find anomalies in the data.
In the journalistic process, AI adds value at every stage: from research and discovery through production, all the way to post-publication audience interaction and developing follow-up stories. AI also helps us reach audiences who might have had trouble accessing reliable information in the past — say, because of language barriers or physical limitations. Think of translations and subtitles. And for audiences more broadly, AI enables seamless packaging of news for different needs: one reader wants deeper background info on a topic, another just wants tight bullet points in plain language.
At the same time, AI brings new challenges, such as verification. The importance of visual fact-checking in photo and video journalism grows every day. As journalists, we need to stay sharp and learn new tricks to keep pace with the latest advancements.
You are building an AI tool for Yle’s newsrooms: can you give us a glimpse of what it is and how it can support journalists?
Jyri Kivimäki: When we started exploring generative AI in journalism, we wanted every single person at Yle to be able to learn and understand this technology hands-on, in a safe environment. That’s why we built our own interface – YleGPT – and we have been deliberately developing it together with journalists into a tool that adapts to different needs.
With YleGPT, you can transcribe your interview recordings and use various AI assistants powered by language models – a drafting assistant, a news wire assistant, one that turns podcasts into text articles, an archive search assistant, and many more. On top of that, YleGPT is a gateway to new AI possibilities, including tools that journalists themselves have vibe-coded for specific needs. A great example is a text-based audio editor that one of our regional reporters coded in no time for radio and podcast work.
If you were to sum it up, what do you view as the main risks and the biggest opportunities of introducing AI into journalistic practice?
Jyri Kivimäki: Right now, in many large organisations AI is still very much tied to individual tools and personal workflows. It is not yet a structural part of how newsrooms operate, which can easily lead to inequality within teams. One reporter might be using AI at every stage of their work, while the colleague sitting next to them still handles AI-friendly routine tasks manually. That makes newsroom management tricky. The AI-empowered journalist can focus on producing more meaningful content, while the one without AI is stuck in old ways. It’s critically important to support learning and give people real opportunities to adopt new ways of working. In the end, the change isn’t really about the technology, it’s about how we work.
What can participants expect from your workshop at the 2026 European Science-Media Hub Summer School?
Jyri Kivimäki: I’ll be sharing practical insight from our work at Yle: very concrete and hands-on examples of AI use for news discovery and data mining, cross-platform versioning, and vibe coding for storytelling tools and interactive journalism. I will try to bring my best to the table, and since this field moves so fast, there will probably be new themes that weren’t yet on my radar when I submitted my workshop proposal.
Responsible use of AI is very close to my heart, so the ethical dimension is embedded throughout the workshop – not as a separate lecture, but woven into everything we will do during those three hours.