Europe and Denmark must reduce dependency on external digital innovation. This requires people who can speak both the language of computing and the language of their field – what we call scientific bilingualism.
It sounds simple, but in practice it is complex. Universities and industry must work more closely together to create frameworks that support this educational transformation.
This was the core message from Jan Madsen, Professor and Head of DTU Compute, at the Digital Tech Summit – a conference where universities and industry meet to discuss and explore the many facets of digitalisation.
Computing is more than just AI
Before a panel debate, Jan Madsen introduced the concept of scientific bilingualism. First, he explained what computing really means:
“Digitalisation is about how we can use computing in a very broad way. I deliberately use the term computing and not just AI, because computing is a much broader concept. AI is the buzzword of the moment, but there are many other aspects that are also important – in fact, even for AI.”
Computing covers both breadth and the entire technology stack – everything from cloud data centres to edge devices, from mobile phones to the smallest sensors and actuators. Common to all is that they are driven by chips, the engines that perform computation and data processing.
“On top of that we have programming, which can include AI algorithms, and it is important to understand that these algorithms exist both in large data centres and in very small devices. To handle security, energy and speed, we can also build chips dedicated to AI algorithms.”
“And whenever we communicate across platforms, we need well-defined cybersecurity. So, when we talk about computing, it is a broad field where AI is an important component, but not the only one.”
Communication gap between disciplines
Scientific bilingualism spans several steps.
“It could be within science, construction, biotechnology – where companies hire graduates for specific tasks, and where knowledge is kept up to date through the way we conduct research,” Jan Madsen said.
Traditionally, research has been an interplay between theory and experiments, where new models must be validated in practice, or where we try to understand what happens in the lab and build models accordingly.
Industry increasingly needs people with computing skills, which is a full discipline in itself at universities.
“The challenge is that if you just hire the best data scientists, they can be super-efficient, but there may be a gap to other domains. Innovation does not happen simply by putting data scientists and biologists in the same room. We need to challenge this gap, and one way to do that is to create programmes that combine clear elements from both computing and the domain in question. The goal is for graduates to have competences in at least two domains. That is what we mean by scientific bilingualism,” he said.
If this is to work well, universities could create joint programmes and develop integrated courses so that graduates become even better at working in the space between disciplines. In this way, Denmark could expand the computing paradigm to permeate all areas and transform them.
But as mentioned earlier, this is not easy.
Three competence profiles
Jan Madsen described three competence profiles.
We need deep specialists (I-shaped profiles), but if we do not work with computing, we risk missing opportunities.
Many today work as T-shaped profiles – with depth in one area and some knowledge of computing as a tool. Here we risk being locked into incremental development and missing major breakthroughs. Computing is not just a tool; it is a mindset.
The ideal profile is π-shaped; two deep specialisations combined with broad understanding – where computing is embedded in the discipline – bridging technology and domain knowledge.
(Competence profiles: I-shaped = deep specialist; T-shaped = depth + some breadth; π-shaped = two deep specialisations + broad understanding).
The story continues below the photo.
