Klaus Beetz, ceo van EIT Raw Materials tijdens de Research & Innovation Days
Author profile picture

Half of Europe’s small and medium-sized enterprises (SMEs) in the raw materials market are afraid that they will be left behind because they do not have the capabilities to utilize artificial intelligence. This is evident from a study conducted by the European Commission’s investment vehicle for raw materials, EIT RawMaterials, based in Paris.

This is likely to concern more than one million companies in total. The EU has 2.1 million companies active in the field of raw materials, according to figures presented by the director of EIT RawMaterials, Klaus Beetz, to the European Commission during the recent Research & Innovation Days. These provide 32 million jobs, 13 million of which are in the high-tech industry. Based on figures from 2016, they account for 16% of the European Gross Domestic Product.

AI can preserve raw materials

The utilization of AI in the raw materials industry leads to reductions in the use of raw materials and materials in general. But AI can also make manufacturing processes more efficient and can improve maintenance of machines, ensuring that they last longer. Therefore, it is a shame that leaders of half of these companies do not think that they are able to make use of AI.

In effect, they remain stuck in a situation where, for example, a factory operator has to be able to listen in order to hear whether a machine needs to be lubricated, even though this can now be done automatically. “I remember buying my first car,” says Beetz. “I didn’t have much money. So I ended up buying an old car. When I was driving, I used to anxiously listen to engine noises that might have been a sign of a fault. Sadly, that was often the case!” he laughs.

That classic bucket of bolts is long gone for most people. These days, lights in a car switch on automatically when a part becomes too hot, for instance. Or when it uses more energy than it should. AI can analyze what is going on based on similar experiences from other cars, to name one example. Signals are then automatically sent that notify you that a certain part requires some maintenance.

Not enough AI professionals in the healthcare sector

One of the reasons why half of the SMEs working in raw materials are afraid of falling behind is probably attributable to the lack of personnel who are trained in AI and are capable of incorporating it into companies.

This is a major problem in medical care. This is what the Director of EIT Health, Jan Philipp Beck, has also mentioned. EIT Health invests in innovative start-ups in the healthcare sector. One of the solutions that Beck came up with is to integrate AI into the medical training of future physicians.

Another problem that arises with the use of AI in healthcare is that it may violate patient privacy laws. Consider, for instance, the use of a robot that takes over the work of a hospital nurse. This robot distributes medication to patients, responds to readings of physiological functions, and brings food around. This can then be communicated to the platform on which the patient’s file is located. It can even signal whether the patient is e.g. smoking or drinking. All this combined, a system like this would generate a lot of personal data that could be linked to the patient’s condition and recovery.

Problems with privacy laws

Pretty handy, you would think. But these should not fall into the hands of any party other than the hospital in question. However, the hospital may also want to use this data to do research aimed at creating AI that can predict the course of a disease.

This problem doesn’t arise at all when AI is used for managing the behavior and maintenance of material and equipment, Beetz said. As such, it really is a missed opportunity that so many companies are not using AI in this sector.

There has to be trust if you want AI to be widely applied in healthcare, Beck pointed out. But how can you trust a system when you have no idea how it works? – That was one of the questions posed by a listener.

Well, yes. A programmer is merely a human being, of course. The director of EIT Digital in Eindhoven, Willem Jonker, explained that decisions that you make as a human being are based on certain convictions. The same applies to a programmer who creates an algorithm. But how do you know whether that programmer has programmed those decisions honestly? The answer to that is, as an outsider of an organization that is making use of an algorithm, you usually are in the dark about that.

Algorithms need to be transparent

“It is really important to be transparent about the choices the algorithm makes,” Jonker said. A programmer of an algorithm should also be monitored by a supervisor. This ought to prevent decisions based on preconceptions.

Albeit, there is no generally accepted standard for this at the moment.