The Dutch government published its national plan for artificial intelligence on Nov. 6. It sets aside €120 million for the Netherlands' part in a joint European AI project, to which Dutch companies and research institutes can contribute by teaming up with partners in other countries. It also reports on a study already under way on whether the country should set up its own AI safety institute, and says the study's advisory report is expected to reach parliament in December 2026.
The ministers approved the plan on the proposal of Willemijn Aerdts, the junior minister (state secretary) for the digital economy and sovereignty, who coordinates the work. The government says the goal is not complete independence from foreign technology. It wants enough knowledge and capacity of its own to keep access to important technology and to keep making its own choices.
The €120 million is for the Netherlands' part in what EU rules call an Important Project of Common European Interest on AI. Under those rules, the European Commission can approve government support for a large project shared by at least four EU countries that carries high technical or financial risk. The Dutch plan says the aim is to develop new AI technology, infrastructure and uses. It also wants to strengthen the Dutch and European AI industry, the country's control over key technology and its ability to compete.
The government is modeling the safety institute idea on bodies abroad such as Britain's AI Security Institute, a government-backed body that studies what advanced AI systems can do and what effects they have, and develops and tests ways to reduce their risks. Five Dutch ministries, including the defence and justice ministries, are involved in the study. The government says such an institute could help the country spot new risks and deal with them in time. The study is of a Dutch institute, though at one point the plan says a Dutch or European institute could help the country steer fast-moving AI.
The plan points to a 2024 report by the Dutch intelligence services and the national counterterrorism coordinator, which found that AI will probably strengthen nearly every existing threat to the country. The report says AI could lead to more or better-executed sabotage, disinformation, hacking, radicalization or theft.
The plan also promises more plans. In the first three months of 2027, the government will send parliament a letter on AI infrastructure, including the computing power needed to build and run AI systems, and another on how the government will steer the growth of data centers. The government wants to give companies room to try new AI uses under the supervision of regulators it has yet to name, and it plans a pact with the media industry that will cover how AI systems use news content.
The plan identifies the project only as a joint European project on AI and does not say which countries or companies take part, or when the money will be spent, though a news report says it becomes available from 2027. The plan gives no cost or timetable for a safety institute.