40 Years After the Palme Assassination: LNU Researchers Bring New Insights With AI
On February 28, it will be 40 years since Sweden’s Prime Minister Olof Palme was shot and killed in the street. Despite decades of investigation, thousands of interviews, and extensive national and international leads, one central question remains: Who killed Olof Palme?
At the Department of Computer Science and Media Technology at Linnaeus University, two researchers, Mattias Davidsson and Tibo Bruneel, have in recent years been working to analyze one of the world’s most extensive criminal investigations. Through both traditional research and advanced AI technology, they aim to shed new light on one of Sweden’s greatest mysteries.
LNU Researcher Featured in New TV4 Documentary
On the anniversary of the assassination, February 28, TV4 will release a new documentary featuring researcher Mattias Davidsson. His interest in the case began in 2018—somewhat unexpectedly—when he was searching for a sleep-friendly podcast.
“It turned out to be somewhat counterproductive when it came to actually falling asleep,” he says.
When the investigation was officially closed in 2020, Mattias encountered the Palme Murder Archive, a digital repository containing released documents from the investigation. This led to deeper insight into the vast body of historical accounts, documents, and theories surrounding the case.
I soon realized how much contemporary history and Cold War narratives are hidden in the material, closely connected to the Palme assassination and now highly relevant again. Secret resistance movements, Stay Behind networks, disinformation—stories that often sound more fantastical than any Bond film.
Over time, technological developments have opened new possibilities. Today, the group increasingly uses advanced AI tools to organize, visualize, and identify connections within the enormous volume of documents.
“What I personally find most interesting is how AI can help transform evidence into structured arguments and assess the strength and relevance of those arguments for or against different hypotheses,” Mattias concludes.
Industrial PhD Candidate Tibo Bruneel Develops AI Solutions for Digital Forensics
In parallel, Tibo Bruneel, one of our industrial PhD at Linnaeus University and active at Softwerk AB, has approached the Palme material from a more technical perspective. In collaboration with the well-known podcast Spår, he has analyzed and structured large volumes of data related to the investigation.
The central question is: how can AI be used to identify previously undiscovered clues in an archive spanning four decades?
This work resulted in the development of PalmeNet-Chat, an AI-based tool that enables users to search, reason, and analyze the archive in an entirely new way.
Digital forensics is no longer just about retrieving data; it’s about making it usable. With PalmeNet-Chat, we don’t just search old PDFs. We provide investigators with a system that can actively cross-reference and analyze decades of complex investigative material.
Softwerk has also collaborated with the Belgian AI company ML6 to develop a cloud-based version incorporating Deep Research and agent-based language models, enabling even more advanced analysis of large document collections.
The Palme assassination is not only a criminal investigation; it is a significant chapter in modern Swedish history. The combination of human expertise, archival work, and advanced AI makes it possible to shed new light on both old and new leads. The work of Mattias Davidsson and Tibo Bruneel demonstrates how academia, industry, and digital tools together can contribute to understanding one of Sweden’s most intricate and debated cases—40 years later.
More information and links
The documentary featuring Mattias Davidsson will be released on February 28 on TV4. Watch it on this webpage.
Would you like to learn more about Tibo Bruneel’s research on the Palme case? You are welcome to attend the doctoral seminar at Linnaeus University, where he will present his results on Friday, March 13, 2:00–4:00 PM. More information and registration are available on the university’s website.