What happens when you build an AI hub inside a newsroom?
For Aftonbladet, Scandinavia's largest newsroom, it meant creating tools that streamline editorial work and transform reader engagement. From AI-generated article summaries to an election chatbot that handled 18,000 questions on its first day, this dedicated team of editors, developers, and designers has been experimenting with tools that enhance how news is produced and consumed.
In this episode of Newsroom Robots, I spoke with Martin Schori, Deputy Managing Editor, Associate Publisher at Aftonbladet, and Program Lead for the AI Hub. We discussed the AI Hub's biggest successes, challenges, and the valuable lessons learned along the way.
Here are three key takeaways from my conversation with Schori:
1️⃣ AI Summaries Boost Reader Engagement: One of their most successful tools, AI-generated summaries, has significantly increased reader engagement. Schori explained that 40% of readers who interacted with the summaries spent more time on the articles.
2️⃣ Election Chatbot's Success: The EU election chatbot was another standout project, answering about 150,000 questions from readers, with 60% being unique questions asked by users. A key highlight of the chatbot was its ability to significantly boost user logins, a persistent challenge for news media in Scandinavia. Typically, login success for such initiatives hovers around 5%, but this project saw nearly ten times that rate, making it a major win for Aftonbladet.
3️⃣ Tracking Diversity with AI: Aftonbladet used AI to analyze a year’s worth of content—around 100,000 articles, along with thousands of social media posts, videos, and podcasts—to uncover gender and diversity gaps. The analysis revealed a significant gender imbalance across different departments, with some areas showing a larger gap than expected. While certain coverage, such as political or sports reporting, is inherently male-dominated due to the nature of the stories, other areas provided opportunities for improvement. AI tools also helped identify female experts to balance representation, and a bot was developed to monitor the gender balance on the front page. Although not perfect, these AI-driven tools are helping Aftonbladet actively work on closing the representation gap in their content.
The Pivot: From using AI for editorial tasks to building audience-facing AI products
The most striking aspect I took away from Aftonbladet's AI journey is how they adapted their strategy based on their findings. Despite initial excitement, they found that AI wasn't as well-suited for many editorial tasks as hoped. This realization and lower-than-expected adoption rates for their prototyped editorial tools led to a significant shift in their AI strategy, prompting the team to refocus their AI efforts on audience-facing products and experiences.
With nearly a year of hands-on experimentation, Aftonbladet has now accumulated invaluable insights into what works, what doesn't, and why.
Aftonbladet's AI journey offers crucial lessons. It highlights the importance of being willing to experiment boldly, fail productively, and pivot decisively based on outcomes.
Their journey shows that successful AI adoption isn't about forcing the technology into every aspect of newsroom operations but about experimenting and finding the sweet spots where AI can genuinely enhance journalism and audience engagement.
Curious to learn more about Aftonbladet's AI experiments? 🎧 Listen to the full conversation available now on Apple, Spotify, Google Podcasts, and other major podcast platforms.
The AI Hub Powering Innovation at Sweden’s Aftonbladet: In Conversation with Martin Schori