To prevent the Netherlands from lagging behind in the development of artificial intelligence, a large-scale national investment plan must be drawn up, comparable to the Delta Works. This requires new legislation, billions of euros to be invested annually and new research institutes to be established.

More than fifty Dutch AI experts write this in the National AI Delta Plan that will be published on Monday. The experts include entrepreneurs, scientists and experts from the media and education sectors, who work with artificial intelligence every day. The plan was written at the request of outgoing Minister of Economic Affairs Vincent Karremans (VVD). According to the minister, the new plan offers “very good, concrete points that the government can work on,” he said in a telephone response.

According to the AI ​​experts, the Netherlands is facing a “historic moment of choice”, where, if no action is taken, the country risks becoming dependent on Chinese and American AI technology. “AI will become the engine of productivity, innovation and competitiveness in every sector. The coming years will decide whether the Netherlands remains an independent player or becomes a user in someone else’s ecosystem,” the experts write.

According to the experts, the Netherlands has all the ingredients to be a leader in AI technology: chip machine manufacturer ASML, top universities and strong sectors such as high-tech, agriculture, logistics and energy. What is missing, according to the experts, is a “national effort aimed at a shared goal.”

The new cabinet must play a leading role in this, including by appointing a State Secretary AI who reports to the Minister of Economic Affairs. Karremans is in favor of that, he says. “I think a State Secretary for AI is a very good idea, because it is clearly invested somewhere.”

More AI factories

If the Netherlands wants to remain independent of American and Chinese technology, it must invest heavily in its own AI infrastructure, according to experts. Think of data centers and thousands of GPUs – the graphics processors needed for AI. Now the fastest computer in the Netherlands – ‘Snellius’ in Amsterdam – has 640 of these GPUs. A new AI factory in Groningen will soon be added, costing 200 million euros, with approximately 2,500 GPUs.

Also read

Groningen is building an AI factory: how Europe wants to break away from the American tech giants

For comparison, the American tech company OpenAI already owns more than a million of these types of chips. If the Netherlands wants to come close to this, it must set up a European AI infrastructure together with other European countries. And, the experts say: The Netherlands must solve pressing problems surrounding the overloaded energy grid and nitrogen policy, so that more AI factories can be built.

Many proposals are about investing in popular adoption of AI, by setting up education programs and giving civil servants access to the best models. Legislation and regulations also need to be amended. For example, the government must help limit the financial risks of venture capitalists and pension funds that want to invest in the tech sector as much as possible.

It should be easier for startups to fire employees who earn high salaries, so that young companies dare to take more risks to attract top talent. The experts argue for a ‘special economic zone’ with more room for experiment, whereby “it can be temporarily deviated from existing laws and regulations”.

One of the examples of this: more experiments with self-driving cars on the road, something that is currently not allowed in the Netherlands. According to the experts, the Netherlands must designate a city where self-driving cars and delivery drones will eventually be “widely deployed”. Self-driving taxis are currently driving around in many American and Chinese cities, but Europe is still reluctant due to regulations and doubts about safety.

Also read

Good morning, says the car without a driver. Will the self-driving car make a breakthrough?

A self-driving Jaguar taxi from tech company Waymo drives through Los Angeles.

‘Impossible breakthroughs’

The question is how feasible the plans are. The Dutch tech sector has become disillusioned with politics in recent years, partly due to the constant change of responsible ministers. AI was hardly a topic in the last election campaign.

Yet the inventors of the plan are optimistic, say initiators Michiel Bakker – assistant professor in AI at the Massachusetts Institute of Technology – and Jelle Prins, founder of the Amsterdam AI company Cradle. “Investing in AI is not a polarized issue. We have the impression that parties from left to right believe that we should do something about this,” says Bakker. “The question is where we put the emphasis.”

Prins hopes that politicians “see the plans as an investment and not as a cost item,” he says. “If we do nothing, we will lose our competitive position. Soon all our companies will use American AI models, until Trump suddenly says that we are no longer allowed to use them.”

The most far-reaching (and therefore most expensive) proposals from the experts concern a number of ‘flagships’ that the Netherlands should establish. Such as an AI Impact Institute, where Dutch research into the social, economic and security impact of artificial intelligence is bundled.

And the most ambitious plan: a Dutch research center that will achieve ‘impossible breakthroughs’, the National Agency for Disruptive Innovation (NADI). Earlier this year, Karremans announced that he wanted to investigate the introduction of such an institute.

The NADI must become a place where participants receive a budget from the government to develop something with a major social impact within five years, while this is considered technically or organizationally impossible. Consider building a house within a few weeks to solve the housing crisis.

The idea is inspired by the American research institute Darpa, which was the basis for major technological breakthroughs such as GPS and stealth technology for aircraft. By giving researchers complete autonomy, Darpa spawned perhaps the single most important invention of the last century: the Internet.





ttn-32