Netherlands considering AI security agency to address serious concerns
Prime Minister Rob Jetten wants the Netherlands to follow other countries in establishing a safety institute for artificial intelligence (AI). He reacted positively to a proposal made by Volt leader Laurens Dassen. Jetten would like to set it up with countries like France, the United Kingdom, and Germany, nations that already have such an institute.
Jetten said the Netherlands plans to work with like-minded countries next week during the United Nations General Assembly to establish stricter rules and safeguards for AI. The group will include EU member states, while AI powers China and the United States will not take part.
PVV leader Geert Wilders asked Jetten whether the Netherlands is capable of “shooting back” if another country uses AI for an attack. Wilders sees no point in international agreements. “The Chinese and the Russians cannot be trusted.”
Jetten said that Dutch safety services are already among the best in the world and that investments are being made to acquire the best people. The Cabinet is spending tens of millions of euros in the coming years for the AIVD, the Dutch intelligence service.
Dutch and Belgian media companies are working together on guidelines for more transparency regarding the use of AI in audio and video content. NPO, DPG Media, Mediahuis, and Talpa are working on a collective approach at the initiative of Media Campus NL. A spokesperson for Media Campus NL said guidelines for the use of AI in text will also be discussed “soon.”
The guidelines were drawn up in response to the European AI Act, which requires media organizations to disclose certain uses of AI, including deepfakes. They will help creators identify what counts as a deepfake, determine when AI-generated content must carry an AI label, and establish how long that label should remain visible.
In this, there will be different rules for different types of content, with news and journalism treated differently from entertainment, reality TV, and fiction. AI use does not always have to be disclosed. Technical applications that leave the content, meaning, or context unchanged generally do not require an AI label, according to the guidelines.
The guidelines, which can be found on Media Campus NL’s website, will be developed further in the coming period and altered based on experiences and feedback.
Media Campus NL is an organization that stimulates innovation and cooperation within the Dutch media sector.