KTH Logo

KTH takes a holistic approach to AI

Anders Söderholm

 , 2026-11-11

It has been said many times: AI is changing everything! It is changing the way we conduct research, teach and run our operations, and this change is happening faster than most of us thought just a year ago. New technology requires clear guidelines. KTH is therefore introducing principles for the use of AI, while also establishing an Advisory Committee to strengthen our work in this area.

For a technical university, AI represents a fundamental transformation. We are helping to lead the way in this area through our research and by educating the engineers who drive AI development in industry. At the same time, we are facing pressure for change in virtually all areas.

KTH has long been building a strong foundation in AI research, covering areas such as machine learning, robotics and autonomous systems. Since 2015, we have been one of the partner universities in WASP (Wallenberg AI, Autonomous Systems and Software Program), Sweden’s largest research programme in this field. KTH also coordinates the Ellis Unit Sweden and is involved in major programmes such as Digital Futures, SeRC (Swedish e-Science Research Centre), DDLS (Data-Driven Life Science) and Cybercampus Sweden, all of which feature AI as a central and significant component.

We have also brought together research in the form of a strategic AI initiative with three focal areas: the fundamentals of AI; AI for scientific discovery; and AI for technology and society. Research that truly makes a difference often occurs at the intersection of AI and other fields, such as materials science, energy, the life sciences, transportation and urban planning. Therefore, we need to bring together the full breadth of AI, ranging from basic research to industrial applications.

Without computing power, there can be no competitive AI research.  KTH’s computing capacity is a flagship resource within the national NAISS (National Academic Infrastructure for Supercomputers in Sweden) infrastructure and is used by researchers throughout the country. The PDC (Parallel Computing Centre), now part of the KCSC (KTH Center for Scientific Computing), is a key NAISS partner. Computing capacity is now as important a part of the research infrastructure as laboratories and advanced instruments, and these types of resources require the expert support provided by the KCSC.

AI is already integrated into our programmes, ranging from civil engineering and computer science to electrical engineering and machine learning master’s programmes. However, it is essential that all of our engineers, architects and teachers understand AI, both as a tool and as a force for social change. That is why we are now reviewing our entire educational portfolio and scrutinising the content to ensure that we are building the right competencies with our students. This is taking place at both the course and programme levels.

Guidelines and recommendations regarding the use of AI in teaching have been in place for some time. We have now established principles for the use of AI across KTH as a whole, and these will be published shortly. These principles are based on taking responsibility for our use of AI, complying with the law, acting ethically in line with KTH’s core values, and creating conditions that allow for agility, rapid development and impact. People must always be kept informed, and the use of AI must be transparent.

At the same time, AI touches nearly everything we do today, from research and education to infrastructure, administration and collaboration. This is why we are establishing a special AI Advisory Committee. The purpose is to compile issues, provide a holistic perspective and lay the groundwork for decisions, ensuring that our initiatives reinforce one another rather than moving in different directions. We need to act swiftly, as the reality is changing rapidly, and the work group will enhance our ability to do so.

There are, of course, plenty of issues on the agenda, but we are well equipped. We have the capability and knowledge to handle the AI revolution both responsibly and strategically!

Anders Söderholm avatar
Anders Söderholm

Categories:

Digitalisation, Education, Innovation, Research

Europe is also built on campus

Anders Söderholm

 , 2026-10-28

In mid-October, over 300 participants from nine European universities gathered in Stockholm for this year’s Unite! dialogue, hosted by KTH.

The University Alliances is a flagship initiative launched by French President Emmanuel Macron at Sorbonne University in Paris in 2017. The original idea was to bring together universities across Europe, transcending national, linguistic and cultural boundaries, in order to foster a sense of European identity and cohesion. Later, the ambition to foster European competitiveness and innovation in response to a new geopolitical landscape was added to this vision.

KTH’s European university alliance, Unite!, consists of nine universities spanning several countries, primarily focusing on engineering and the natural sciences. Within Unite!, mobility and development are fostered through extensive student and faculty exchanges, seed funding, and joint projects. The alliance is also founded on a shared set of values and emphasises the importance of European solutions to our various national challenges.

We are far from achieving Macron’s vision of a united Europe. Similarly, we have a long way to go in terms of raising the visibility and accessibility of Unite!’s activities on campus. The alliance needs to facilitate collaboration and increase its value and the visibility of what can be achieved together. This is why this year’s Unite! conference focused on transitioning from projects to more established structures, growing and maturing from an alliance of extraordinary activities to working together in a more organic and integrated way on a continuous and daily basis.

Ultimately, university alliances are an expression of a shared European desire to strengthen the bonds between our countries, bridge our differences and help free our continent from conflict and the risk of war. That is worth fighting for.

Anders Söderholm avatar
Anders Söderholm

Categories:

Co-operation, Education, Innovation, Internationalization, Research

Time for the state to act on AI research

Anders Söderholm

 , 2026-10-14

In an article published in Dagens Industri on 30 September, Björn Rosengren argues that the government has abdicated its responsibility for AI research. This is an important and clear contribution to the debate on AI. KTH is ready to take responsibility for its part.

The debate on AI has become increasingly heated recently, with a wide range of comments on the necessary actions in various sectors and the responsibility of policymakers in addressing this fundamental societal issue. Rosengren highlights something that we at the universities have recognised for a long time. The Swedish AI debate focuses on data centres and energy, the need for venture capital, scaling up new business ideas and reforming primary and secondary education. All of this is necessary. However, the latest AI breakthroughs in fields such as mathematics demonstrate how quickly the conditions for research itself are evolving. Without world-class research and the people it trains, the infrastructure becomes empty and has no future.

New methods emerge and the next generation of engineers and researchers is shaped in research environments. It is also there that businesses and society can access the expertise needed to use AI wisely and responsibly. Unlike many other stakeholders, universities are not merely subject to the AI revolution. We have the capacity to create, drive and steer the development that is now unfolding at an ever-increasing pace.

Sweden has been fortunate in many ways. Private initiatives, not least the investments of the Wallenberg Foundations, have laid a strong foundation. However, private forces cannot fully replace the government’s long-term responsibility. The countries we compare ourselves to are now investing vigorously and strategically. If we do not match their levels of investment, we will lose both researchers and competitiveness, and it will take a long time to recover from that loss.

As soon as it takes office, the new government will have an opportunity to demonstrate its leadership. It will require courage to provide long-term and predictable resources for AI research, graduate education and the computational capacity that this research requires. KTH and Sweden’s other universities are ready to play our part. Now, policymakers must take a significant step forward by creating long-term conditions that extend beyond government formations, campaign promises and annual budget decisions.

Anders Söderholm avatar
Anders Söderholm

Categories:

Digitalisation, Politics, Research

Global challenges call for global university partnerships

Anders Söderholm

 , 2026-10-01

This week, I have had the opportunity to visit South Africa and Tanzania to strengthen our ties with some of our African partner universities, and to announce some news ahead of KTH’s anniversary year in 2027.

The Global Development Hub (GDH) is a KTH initiative established ten years ago to develop approaches to tackling various global challenges through challenge-based education and innovation. This innovative approach has also become a fundamental model for our collaboration with a select number of African universities. By partnering with like-minded universities, the GDH can make a real difference.

Our original GDH partners are Strathmore University in Kenya, the University of Dar es Salaam in Tanzania, the University of Rwanda in Rwanda, and Botho University in Botswana. This week, KTH visited the University of Dar es Salaam (UDSM) to gain a better understanding of our ongoing collaborations and discuss future developments. Over the years, there has been increased student and researcher mobility between UDSM and KTH, meaning that we now have a deep understanding of one another and trust each other, which makes it easier for us to collaborate on new development projects.

KTH has previously been involved in ‘sandwich’ programmes aimed at building capacity in postgraduate education at UDSM. These programmes have helped to foster a close relationship between the two institutions. Students from both universities have participated in courses and workshops, and the I-Grid project, which aims to secure the energy supply to end users in Tanzania, has been underway for some time now. The I-Grid project combines researcher exchanges with a challenge-driven approach to problem-solving and technological development.

In South Africa, KTH has visited several universities to strengthen its presence in the country and deepen its links with academic institutions. Both South Africa and several countries in East Africa are experiencing rapid economic development, whilst simultaneously facing major societal challenges that need to be addressed. Universities and university collaborations play an important role in the wider societal transformation taking place. Many issues are also similar to those we face in the Nordic region, such as access to clean water and energy.

The Times Higher Education (THE) World Academic Summit is also being held in Cape Town this week, hosted by the University of Cape Town in collaboration with the University of Bristol. This annual conference brings together university leaders from around the world to discuss important issues.

This year, much of the agenda focused on the new geopolitical world order and how universities around the world are responding to its implications. Topics of discussion included the emergence of AI, the associated threats and opportunities, and academic freedom and institutional autonomy in an increasingly fragmented and polarised political environment, where universities can no longer take trust and confidence for granted as the basis for their engagement with society. Clearly, universities across the globe face many similar challenges, even if the consequences for individual institutions differ.

On the final day of the conference, the host for next year’s THE World Academic Summit was announced: KTH. Hosting this major global conference will be one of several events marking KTH’s 200th anniversary, and I am looking forward to welcoming colleagues from all over the world to Stockholm in October 2027.

Anders Söderholm avatar
Anders Söderholm

Categories:

Education, Innovation, Internationalization, Research

No, engineers are not the nation’s enemies

Anders Söderholm

 , 2026-09-16

Sometimes columnists get it wrong! A recent example can be found in Aftonbladet, where Saga Cavallin in her culture article  ”Här är väljargruppen som ingen pratar om” (Here is the voter group nobody talks about) published on 13 September, uses vacuity as an argument to attribute all manner of evils to engineers.

Cavallin writes that “The engineering mindset of the Swedish national character has given us ball bearings, Bluetooth and dynamite. But it has also been blamed for racial biology, forced sterilisation and the global export of soulless arena EDM. Perhaps it is also partly responsible for four years of the most sadistic, inhumane and anti-democratic politics witnessed on this side of the 1930s.”

This argument is based on two unusually flimsy lines of reasoning. The first is a study of around a hundred people active in anti-media online communities, half of whom were found to have an engineering background. The second is the book Engineers of Jihad, which claims that engineers are over-represented among violent Islamist groups. In the study, around 200 people could be linked to a specific educational background, of whom just under 45 per cent had studied engineering subjects.

Even more serious is the conflation of over-representation with the characteristics of the majority. While it may be true that 50 people in a study of hostility towards the media are over-represented in relation to that group’s share of the labour market, to conclude that this characteristic is possessed by the majority of practising engineers is misleading and incorrect.

It is similar to when it has been noted that immigrant men are over-represented among rapists, and this characteristic is then attributed to all immigrant men. Similarly, when it is said that people with certain mental disabilities are over-represented among those convicted of violent crimes, it is assumed that all, or a significant proportion, of people with these disabilities are potential murderers.

Similarly, it seems somewhat odd to claim that an IT consultant in Mörbylånga is likely to have a hostile attitude towards the media and an engineering-oriented mindset based on a study of Islamist groups that advocate violence in another part of the world. This line of argument is reminiscent of a 31 May 2024 editorial in Dagens Nyheter by Gina Gustavsson, who used the same study that Cavallin is now relying on to claim that there is a link between engineers and white supremacy and neo-Nazism.

Drawing the conclusion that an entire profession suffers from a democratic mindset deficit based on a hundred or so hand-picked interviews and a book on terrorism is precisely the sort of reasoning that is usually labelled populism when applied to immigrants or young people, for example. It involves selective sampling, a simplified explanatory model and a dig at a group that rarely gets a chance to defend itself in the culture pages. To label media-critical right-wing movements as populist while constructing an argument based on the same method is, to say the least, both pointless and tendentious.

While it is not unreasonable to suggest that engineering education shapes a way of thinking – as does all education – it is illogical to single out an entire professional group as enemies of the nation. Moving from this observation to singling out an entire professional group as enemies of the nation requires considerably stronger evidence than has been presented. Until then, it is merely a prejudiced game of pick-up sticks dressed up in an unworthy scientific guise.

Anders Söderholm avatar
Anders Söderholm

Categories:

Education