In the week that a series of AI plans were presented in the Netherlands, Pim de Witte’s name also came up regularly. The 30-year-old Nijmegen resident is the founder of Medal.tv, a platform where gamers share short videos of how they play computer games. A gold mine for AI companies that need data to train their models.
The American tech company OpenAI offered reportedly half a billion dollars on the platform, an amount that De Witte would have rejected. Instead, he started a new start-up at the beginning of this year: General Intuition, which uses the game images to develop its own AI models. He pulled last month 115 million euros, an exceptionally high investment for a company that is barely ten months old.
De Witte lives in New York, but his companies are registered in the Netherlands. The question is whether it will remain that way. Because the Netherlands has an enormous amount of AI talent and a thriving start-up scene, but we often see that companies that really want to grow move abroad. In the tech world, De Witte is often cited as an example of an AI entrepreneur who will leave the Netherlands if politicians do not take action quickly.
This week, the tech sector presented a series of plans with ideas about how the Netherlands can participate in the development of artificial intelligence (AI). This started on Monday with a large-scale national investment plan by more than fifty entrepreneurs and scientists, led by entrepreneur Jelle Prins and AI scientist Michiel Bakker. According to the group, AI will inevitably become the engine of productivity, innovation and competitiveness in every sector. If the Netherlands wants to participate, it needs its own AI infrastructure, a series of new research institutes and more flexible regulations for start-ups.
“Build your own data centers”
It arrived later this week a plan from AIC4NL – a coalition of dozens of organizations – that proposed 5 billion in investments in AI. The national tech investor Invest-NL pleaded in turn for targeted investments in which the Netherlands can excel within five years. Not by building your own ChatGPT, but by investing money in niches such as energy-efficient chips or technology that helps robots to better understand the physical world.
The key point is that all plans involve spending billions of euros, which the government itself will only be able to partially finance. The advice refers to a role for the National Investment Bank, which itself has not gone beyond the idea phase that is only being discussed in The Hague for the time being. The idea is that this can attract billions in investments from the private sector, such as pension funds, for, for example, the necessary investments in AI infrastructure.
Pim de Witte was given all the plans, he said from New York via WhatsApp. “I have read them and they are important,” he says. De Witte especially hopes that the Netherlands will take the step to build its own data centers with AI chips, in order to give its own companies a competitive advantage. “As machines make more and more decisions, you have to be able to control them yourself,” he says. “Otherwise you are not a sovereign country.”
AI State Secretary
It is no coincidence that the plans are being presented now. The tech sector is trying to influence a new cabinet through the proposals. With success: according to Jelle Prins, his proposal was discussed at the formation table this week. “I understand that it is being discussed at the right levels,” he says.
Prins and Bakker benefit from their bond with outgoing Minister of Economic Affairs Vincent Karremans, who himself founded a tech company in 2011 (vacancy site Magnet.me). Around last summer, Karremans and his officials had several meetings with the duo and Karremans asked Prins and Bakker to write an AI vision for the Netherlands. Karremans speaks “the language” of entrepreneurs, says Prins. “He understands the urgency and continuously asked: how can we get this done even faster?”
America won’t wait for us to know if we want AI
One of the proposals immediately received support from Karremans: the appointment of a ‘State Secretary for AI’, based at the Ministry of Economic Affairs. Whoever takes control of the national AI strategy will be crucial for the direction the Netherlands takes. For the time being, Economic Affairs seems to be the most logical place, which indicates that the emphasis will mainly be on economic opportunities and investments.
At the same time, there is a risk that themes such as risk management or protection of professional groups affected by AI remain underexposed. Consider the music and media sectors, which deal with AI companies that have built their language models using copyrighted material without always being paid for it.
In the past year and a half, the discussion about AI in Europe has completely changed: from talk about more rules and risk management, to billion-dollar plans for AI factories and new development labs. According to Prins, there is now “momentum” for large-scale investments and the Netherlands must seize this, he says. He fears a one-sided discussion about risks. “We can feed them endlessly, until we are no longer the strongest in the niches where we as the Netherlands now excel. There is a window and it is slowly closing. America will not wait for us to know if we want AI.”
Critical reactions
Yet not everyone is convinced of the new direction. Prins says that in addition to many enthusiastic responses to his plan, he also received a number of “quite aggressive messages” via social media. “People who think AI is a hype or say: it is a stochastic parrot (a device that spouts coincidences, etc.), which cannot be trusted. I thought we were a bit past that point.”
Dutch people are much more negative about AI than citizens from other countries
It is true that the Dutch are generally remarkably negative about artificial intelligence. Last year concluded the Dutch Data Protection Authority states that only one in three Dutch people believes that AI has more advantages than disadvantages. They fear the loss of human control over machines or fear that AI will lead to mass unemployment. This makes the Dutch much more negative than citizens from other countries.
One of the experts who is concerned is internet pioneer Marleen Stikker, director of Waag Future Lab. She is struck by how this week’s discussion about AI is mainly based on economic opportunities and how tech companies determine the political agenda. “A Silicon Valley dynamic of move fast and break things. The trend is: we entrepreneurs know better and for God’s sake give us as many resources as possible,” says Stikker. “We are moving full speed ahead with AI without knowing what we are doing.”
The Netherlands should approach things completely differently, says Stikker. We should first determine together what we as a society find acceptable and what we do not find acceptable when it comes to artificial intelligence, before we let AI companies do their thing. In other words: first rules, then innovation. And not the other way around.
“Public values are seen by tech companies as an annoying nuisance instead of a prerequisite for innovation,” says Stikker. “The basis of generative AI is language, image, media. It is about meaning, relationships and semantics. Let us first consider together how we want to preserve our language and culture. We don’t want all kinds of important sectors to be eaten up by an AI industry, do we?”
The question that, according to Stikker, is no longer asked at all: is more artificial intelligence actually desirable? “It is widely assumed that this is our future,” says Stikker. “I am not saying that there are no useful applications of artificial intelligence. But we should not relinquish control of our future.”
Call for an active role for politicians
In any case, the call for a more active role for politics is growing. And the authors of the AI plans are not completely isolated from the concerns. For example, AIC4NL insists on the development of a fair AI tech industry that operates within all European standards for security in privacy protection. And Prins and Bakker’s Delta Plan calls for maintaining control over the AI transition and making society resilient.
That plan even calls for a National AI Impact Institute, which should conduct ongoing research and advise on the social, economic and security consequences of AI. The initiators also want the establishment of an AI Citizens’ Council in which “representative groups of citizens are systematically involved in the preparation of major choices about AI.”
The advice of all experts explicitly puts a number of important questions on the table for the forming parties (CDA and D66). Is the Netherlands really going to release money to participate in the AI race? Do they leave the almost unstoppable technological revolution to the market? And to what extent will the new government direct the development of AI in Dutch society and economy?
Also read
Expert advice: a national plan such as the Delta Works is needed to participate in the AI race
NEW: Give this item as a gift
As an NRC subscriber you can subscribe every month 10 items give as a gift to someone without an NRC subscription. The recipient can read the article directly, without a paywall.
Why you can trust NRC