Can Europe create AI that we actually understand?

Artificial intelligence is becoming increasingly important, but it is completely dominated by the United States and China. Leaving the field to foreign powers and large companies may seem risky.

It is also concerning that we do not really understand exactly how artificial intelligence works today, according to Norwegian professor Harald Martens.

It is as if we put data into a black box, and then an answer magically pops out. Why? 

Martens believes that Europe should take a different path, one that provides safer, more cost-effective, and more understandable AI solutions.

Current AI is not good enough in all contexts

“Many of today’s AI systems deliver impressive results, but they do not make us smarter, and we do not always understand how they arrive at their results. This can be a problem when AI is used in critical situations,” says Martens.

For over 50 years, Professor Martens has worked on minimalist and realistic data modelling that can be interpreted and understood. 

His work is based on machine learning methods developed and used in fields other than computer science, without neural networks.

Today he is a retired professor who remains affiliated with NTNU. He also works on technical artificial intelligence at the Trondheim-based company Idletechs.

Professor calls for safer problem solving

He is now raising concerns about developments in modern artificial intelligence.

“We need more reliable problem solving, lower energy consumption, and less of a ‘black box.’ At the same time, we need to understand more about what is happening inside the box," he says. 

"We must also try to learn as much as we can about the world we are living in and the system we are working on. And it must all remain under democratic control, based on our Western European values,” he adds.

Martens believes this is entirely possible.

Company develops an AI we can trust more

The field of AI is currently dominated by large, data-hungry neural networks. 

These so-called ‘black box’ models seem easy to use, but they must be trained on enormous amounts of data. 

This type of machine learning requires large amounts of energy, and the resulting solutions are difficult to interpret and critique.

Martens and his collaborators have developed an alternative they call CIM-ML, or Continuous, Interpretable, Minimalistic Machine Learning. 

So what is it exactly?

Instead of feeding the system as much data as possible and hoping for the best, machine learning can begin with our current understanding of the system being described and the measurement techniques being used, the professor explains.

Models that can explain themselves

The machine learning process then proceeds more or less automatically based on streams of modern measurement data. This is the simplest possible solution, but no simpler – compressed and visually interpretable.

“We have chosen to spend more time on the groundwork to combine domain knowledge with modern measurement data. This gives us models that can explain themselves,” says Martens.

CIM-ML is self-learning, self-expanding, and self-correcting. The system describes both known and unknown variations in data. It can process large streams of data on small computers.

“In a time when algorithms are influencing an increasing number of decisions, perhaps the most important issue is not how fast the systems are, but whether we can trust them,” says Martens. 

That requires combining modern measurement data with interpretable mathematics, he explains.

Large language models allow more room for error

The method is currently aimed at moderately complex systems, not enormous language models such as ChatGPT or DeepSeek.

In those systems, there is often more room for error anyway, according to the professor.

CIM-ML is specifically aimed at professional use in fields such as the process industry and other technical systems.

It is also designed for drone- and satellite-based environmental monitoring and Earth observation.

What these new systems can be used for

• Vibration sensors on turbines to distinguish sound, abnormal noise, and measurement interference – before something goes wrong.

• Thermal video monitoring of a smelting furnace or an engine for oversight and optimisation.

• Multi-channel drone or satellite imaging of oceans and land to identify as many observable phenomena as possible.

The goal is not only to obtain a reliable answer, but also to understand how – and ideally why – you received that particular answer.

“In critical applications, it's important to receive early warnings without many false alarms. It's also crucial that the people using the system understand what's happening,” explains Martens.

Is there a European lane in the AI race?

The AI debate often focuses on who has the largest models and the most data. But Europe can choose to take a different path, according to Martens.

The question is whether Europe will invest in more independent and transparent technology.

Or will Europe remain dependent on global tech giants that operate according to the arbitrary whims of transient heads of state?

“The artificial intelligence we are developing is rooted in a proud Western European tradition of democratic enlightenment. My dream for Idletechs has been to establish high-tech jobs in Norway and Europe. These must be based on secure, simple, ethical and understandable machine learning that also promotes human learning,” says Martens.

Reference: 

Martens, H. A Greener, Safer, and More Understandable AI for Natural Science and Technology, Journal of Chemometrics, 2025. DOI: 10.1002/cem.3643

———

Read the Norwegian version of this article on forskning.no

This content is paid for and presented by NTNU

This content is created by NTNU's communication staff, who use this platform to communicate science and share results from research with the public. NTNU is one of more than 80 owners of ScienceNorway.no. Read more here.

More content from NTNU:

• 

  Doctors and nurses believe their own substance use affects patients

• 

  When the offshore wind turbines fall silent: What happens to the waste?

• 

  “Norway used German soldiers to clear minefields and dump explosives in fjords and lakes"

• 

  Physicists may have found the missing link for quantum computers

• 

  This is how emissions from express ferries can be halved in just five years

• 

  Darwin’s giant plant shows how species adapt