AI collaboration for enhanced preparedness
How can Sweden enhance its national preparedness and defense capabilities through collaboration around AI – and what role does AI Sweden play as a facilitator of such collaboration? These were the key questions when AI Sweden and Lindholmen Science Park met with representatives from five Swedish government agencies.
Tommy Schönberg, responsible for Defense Innovation at AI Sweden, kicking off the day
Among the participants, representing the Swedish Defence University (FHS), the Swedish Defence Materiel Administration (FMV), the Swedish Defence Research Agency (FOI), the Swedish Armed Forces, and the Swedish Civil Contingencies Agency (MSB), many emphasized the need for collaboration, particularly with the private sector.
Jon Gustafsson, Head of section, Critical Communication at MSB, was among those highlighting the importance of collaboration around AI. He describes artificial intelligence as a crucial component in many of MSB's future solutions, including the new national communication system for Sweden's total defense, which is currently under development. The agency is also exploring real-time translation as a way to improve collaboration with other countries.
"As a central hub in Sweden’s civil defense, we recognize that collaboration between industry, research, and government agencies is crucial to meet future demands," says Jon Gustafsson. "By joining AI Sweden, we became part of a platform for knowledge exchange and learning from other organizations that are further along in their AI journey. My colleagues and I attended this meeting to accelerate the benefits of our partnership, and we truly feel that we did," he continues.
Jon Gustafsson
Head of section, Critical Communication at MSB
One of the highlights of the day was a presentation of a concept for a new, sustainable, and collaborative ecosystem for civil-military innovation, developed by AI Sweden, Lindholmen Science Park, and the Swedish Defence University. The scope of the new ecosystem is when the time from idea to implementation is less than a year.
“In Sweden, we excel at the individual components of civil-military innovation in the short development cycle, but we need a more efficient way to connect them into a cohesive whole - in short - an ecosystem. In developing the new concept, we identified a set of paradoxes that make collaboration on requirements among stakeholders more challenging than in other sectors of society,” says Katarina Brud from Lindholmen Science Park, who presented the concept.
Katarina Brud
MobilityXlab, Lindholmen Science Park
Identified Symptoms and Underlying Paradoxes
Several critical obstacles – both symptomatic and systemic – need to be addressed:
Symptoms
- Adapting civilian technology for military use takes too long
- Civilian actors lack insight into military needs and problem definitions
- Parallel solutions are being developed for the same challenges
Paradoxes
- Defense-related problems or needs are too confidential to openly seek help
- “If a company collaborates with an agency on early innovation, the company may then be disqualified from procurement”
- The most crucial people are often the hardest to free up time for projects
- Companies and individuals have little financial incentive to engage in collaborations with agencies
Lars Hedström, Senior Advisor at the Swedish Defence University who participated in the development of the concept, emphasized the need to tap into expertise spread across government agencies and industry.
"Adopting AI Sweden’s way of working, we make us faster and better at leveraging cutting-edge expertise that is spread across different actors in the ecosystem" he says. "By quickly assembling agile teams with highly skilled participants from multiple organizations, we can achieve significant innovation progress in a short time."
Later in the day, Mats Nordlund, Director of AI Labs at AI Sweden, addressed the four paradoxes:
"All of these can be solved. For example, the challenge of balancing secrecy with open innovation can often be tackled by generalizing the technical problem, rather than specifying its exact application or performance needs."
One of the meeting’s speakers was Zenseact's Head of AI, Erik Rosén, who discussed the company’s sensor technology for self-driving cars. His talk prompted a very concrete question from the audience: "Cameras are prohibited at our facilities. Can the cameras of advanced driver-assistance systems be disabled?"
Mats Nordlund noted that this is a prime example of value-driven collaboration across sectors. This time between defense actors and industry:
"This type of challenge fits perfectly within our way of working," he says. "It is well-defined and technical. It shows that in our network of partners, we have those who needs to solve a problem as well as those who may have the solutions.This could be a great opportunity to quickly move forward and create collaborative solutions that can benefit everyone."
Another way to accelerate innovation is by involving students. In AI Sweden's case, this is done through a number of talent programs.
"Results from one of our 2024 summer programs demonstrate that the first stages of our innovation model work well," says Mats Nordlund. "In just a few weeks, a group of students developed a prototype called OceanWatch, which uses AI and publicly available data to detect unusual ship movements around Sweden. Now we need to establish mechanisms for taking such projects to implementation with our industry partners, so that the entire ecosystem can benefit, and the solution can be deployed to use."
Mats Nordlund
Director of AI Labs at AI Sweden
Continue reading
AI Sweden forms an AI & Security Consortium
LeakPro enables collaboration around sensitive data
The Swedish Government's AI Commission submits its final report
Follow us
AI Sweden Newsletter
Sign up for the latest in Swedish AI
AI Sweden is formally hosted by Lindholmen Science Park AB.