ĚÇĐÄÍřŇł°ć

19 December 2025

ĚÇĐÄÍřŇł°ć will host a new national centre aimed at developing resilient AI systems. The funding of SEK 60 million comes from the Swedish Foundation for Strategic Research (SSF). The centre will be named Resilience and Security for Trustworthy AI Systems – Resist, and its director will be LiU Professor Fredrik Heintz.

Porträtt av Fredrik Heintz som sitter i en trappa Photographer: Anna Nilsen
Professor Fredrik Heintz at the Department of Computer and Information Science .

“As security threats increase while we simultaneously build a dependency on AI technology in our society, it becomes increasingly important to ensure that this technology is both secure and resilient against various forms of disruption,” says Fredrik Heintz, professor of computer science at LiU.

The use of artificial intelligence is rapidly expanding across numerous sectors, including energy supply, healthcare, and industry. However, the introduction of new technologies also entails risks, particularly in an uncertain global environment where so-called hybrid attacks are becoming increasingly common.

“I believe it is difficult to anticipate many of the risks in advance. It is only when you are exposed to them that you realise what is actually needed. Perhaps we could not have foreseen just how crucial AI technologies would become,” says Fredrik Heintz.

To identify projects that strengthen the resilience of future AI systems against cyberattacks and technical failures, SSF issued a call for proposals. Among the applications received, Resist was deemed to have the greatest potential to meet the objectives. LiU will host the centre, which will also involve researchers from Lund University, Örebro University, and the research institute RISE.

Collaboration for impact

Within the consortium, collaboration will take place with Ericsson, Saab, Epiroc, Sectra, Siemens Energy, Combient, the Swedish Defence Research Agency, the Swedish Civil Contingencies Agency, Cybercampus Sweden, and WARA-PS – a research arena for public safety within the WASP research programme. Fredrik Heintz sees great strength in broad collaboration:

“Research and technology development are exciting – but we want our research to make a real impact! Considering how vital these technologies are for Sweden, it is absolutely essential that we ensure resilient AI systems both in our critical infrastructure and so that our leading companies are not affected by disruptions.”

Fredrik Heintz argues that both research and industry have much to gain from the collaboration.

“Many of our partners operate in hostile environments, so there is considerable experience and knowledge within these companies that we want to harness. We will also build on work within WARA-PS, the cybersecurity testbed at RISE and Cybercampus. Our vision is to make Sweden a role model in safe and reliable AI, with a holistic perspective on cyber resilience throughout the entire chain of development and use of AI.”

Contact

The New Centre

Latest news from LiU

En grupp människor som sitter vid ett bord framför en folkmassa.

Participants in UN climate meetings want greater focus on implementation

UN climate negotiations are often criticised for moving too slowly. A new study published in Nature Climate Change examines what changes government delegates and other participants at the COP29 UN climate conference would like to see.

Neil Lagali vid utrustning för att undersöka ögonen.

Eye problems after COVID-19 can now be explained

Mild COVID-19 can cause severe and long-lasting eye problems, according to a study from LiU. The study also explains why it has been difficult for sufferers to get help: the abnormal eye behaviour cannot be detected by standard methods.

A woman kneeling down in a garden picking plants.

When soil becomes data – how digitalisation affects knowledge of soil health

How do we know whether soil is healthy? A new study from ĚÇĐÄÍřŇł°ć shows that as knowledge about soil is increasingly translated into digital data, important insights about its biological life and local context risk being overlooked.