Something Fishy in Tromsø

It was an ordinary Tuesday morning at the fish processing plant at Vannøya, a small island just outside Tromsø. The sea was calm, the air crisp. Then, suddenly, the distant sound of helicopter blades broke the silence. Fast-moving boats from the Norwegian Coast Guard raced toward the docks. Officers in red uniforms hurried ashore, carrying clear plastic boxes. Within moments, police and inspectors from Norway’s white-collar crime unit, tax authorities, and Fisheries Directorate swarmed the pier. Fishermen stood watching, confused and concerned. Documents and computers were quickly seized from inside the building. Nobody was allowed to leave. Nearby, the Coast Guard vessel KV Barentshav loomed quietly in the bay – an unmistakable sign that this was no ordinary inspection. Across Norway, journalists began scrambling for details. What was going on? Who was involved? Where exactly was this happening?

But at the small newsroom of iTromsø, there was no confusion. They knew exactly where to go. A reporter was already driving fast to reach the last ferry of the day, so they could reach the island. The journalist arrived just as law enforcement had begun looking for evidence of wrong-doing, carrying out files, documents, and computers from the offices on the docks. Cameras clicked, questions flew, details emerged.

iTromsø was the only news outlet on the scene — not because they got lucky. A year earlier, together with journalists from Sunnmørsposten and researchers at OsloMet (Oslo Metropolitan University), they had started to work on a massive database of millions of catch reports, price statistics, and delivery slips. The aim had been to chart sustainability in the Norwegian fisheries, but while applying AI and machine learning techniques to the data, red flags started to emerge Following the events at Vannøya, iTromsø continued to invest in AI, and last year they launched the result of that work: DJINN. We spoke with Lars Adrian Giske, who led the development together with a dedicated team of journalists and developers at iTromsø, Visito and IBM. As the public face of the project, he helped guide the newsroom into a new era of journalism powered by technology.

Creating DJINN

At iTromsø, Lars Adrian was given the task of exploring how AI could help journalists. Together with IBM, and the IBM partner Visito, he started developing DJINN—a computer tool built specifically for journalists. DJINN had a simple mission: it would go through thousands of public documents every day. These could be zoning applications, urban planning proposals, or emails sent between people and the city government.

DJINN’s task was to

• Spot documents that might be newsworthy 
• Summarize them clearly 
• Send instant notifications directly to reporters. 

Suddenly, hours of painful reading turned into minutes of clear results. The DJINN team knew they couldn’t rely on technology alone. He wanted the journalists deeply involved. From day one, they joined in, giving ideas, feedback, and suggestions. DJINN wasn’t just built for journalists—it was built by journalists.

Immediate success at iTromsø

DJINN started working at iTromsø in 2023. Results appeared fast. One young journalist straight out of journalism school soon wrote a front-page story with DJINN’s help—something previously impossible. Before, young journalists spent days struggling to find news. Now they found strong stories right away. 

Journalists now had more time. Instead of spending hours reading boring documents, they could talk to real people. Stories became more human and emotional. 

Lars Adrian gives a clear example: Before DJINN, a journalist might quickly write: “A neighbor complains about losing their view due to new buildings.” Now, they had time to speak with Leif, a 60-year- old man whose living room suddenly faced a huge wall, blocking his view of the mountains surrounding Tromsø. Stories got more emotional and powerful, and readers loved them.

Expanding across Norway

After its success at iTromsø, DJINN caught the attention of Polaris Media—the large media group that owns newspapers across Norway. Polaris quickly wanted DJINN at their other newsrooms. But scaling DJINN was not easy. Each town is different. Tromsø isn’t Trondheim, and Trondheim isn’t Oslo. A story important in the north, like reindeer sightings, wouldn’t be news in southern cities. DJINN needed to adapt. The solution: create two kinds of AI models. One model learned newsworthiness from all journalists across Norway. 

Another learned from journalists locally. DJINN combined these two models, understanding both general and very local news. After this, journalists all over Norway quickly trusted DJINN.

A future of stories we can’t yet imagine

But DJINN was just the start. Lars Adrian already looks to the future. He imagines AI helping journalists in many other ways. DJINN could soon track relationships between people, places, and businesses. It might show hidden connections, revealing conflicts of interest or secret deals. Lars Adrian explained that DJINN might uncover networks that journalists would never find alone—such as politicians secretly involved in companies or hidden influences affecting city decisions. This would help journalists tell powerful stories people couldn’t imagine. He also thinks DJINN could help journalists find hidden problems in health, crime, or sports. For example, data might reveal sudden increases in illness outbreaks, hidden crime waves, or undiscovered sports talents. With AI’s help, journalists would report these stories early and effectively.

Journalism still human

Lars Adrian never forgets something important: computers alone are not journalism. Real journalism, he says, relies on human judgment. It’s journalists who decide what’s ethical and what’s worth telling. Technology helps them make those decisions faster and smarter. This is why Lars Adrian believes journalism is stronger with AI. DJINN doesn’t replace reporters, it makes them better. It lets them find more interesting stories, spend more time talking to real people, and uncover truths hidden in data.

Leading journalism into a new era

Journalism is changing fast. Lars Adrian thinks we must change too. But one thing never changes, he says: the importance of honest, accurate stories. Technology doesn’t threaten journalism, Lars Adrian insists. Instead, tools like DJINN give journalists new power. They help small newsrooms become stronger, faster, and more competitive. In Tromsø, DJINN showed how a small, northern town could lead a big change. That quiet morning, as the police raided the fish plant, something became clear: Journalism had entered a new era. And a modest philosopher-turned-journalist from Tromsø was quietly guiding the way.

From philosopher to journalist

Lars Adrian never dreamed of being a journalist. As a child, visiting Greece at age seven, he told his mother he wanted to become a philosopher. And he almost did. He studied philosophy, taught it, and felt comfortable with big questions about life and truth. But one day, a friend encouraged him to try journalism. Lars Adrian gave it a chance and discovered something surprising: journalism felt very much like philosophy. Both required careful thinking and a sharp eye for truth. He soon became an investigative journalist. This meant spending hours, sometimes days, going through piles of documents. Asking municipalities for reports, filing public requests, waiting, getting rejected, appealing, and finally getting piles of papers. He quickly saw this was slow and painful.