Netherlands rejects AI ban, says Superintelligent AI cannot yet be proven possible

GREAT BAY--The Dutch government has rejected calls to ban the development of superintelligent artificial intelligence or temporarily halt work on the world's most powerful AI systems, despite acknowledging growing concerns that the technology could eventually become difficult to control and pose serious risks to society.

In written responses to Parliament on October 26, State Secretary for Digital Economy and Sovereignty Willemijn Aerdts said the government sees no scientific evidence that superintelligent AI can currently be developed. However, she acknowledged warnings from researchers that artificial intelligence is advancing faster than the safety systems designed to control it.

The government's position follows questions submitted by Member of Parliament Chris Stoffer of the Reformed Political Party (SGP), prompted by warnings from researchers associated with leading AI companies, including OpenAI and Anthropic.

The researchers have raised concerns that increasingly powerful AI systems could eventually operate beyond effective human control, potentially creating risks that governments and technology companies are not adequately prepared to manage.

What Is Superintelligent AI?

Superintelligent AI refers to a hypothetical form of artificial intelligence that would surpass human intelligence across virtually every intellectual field, including scientific research, problem-solving, strategic planning and technological innovation.

Today's AI systems, including popular assistants such as ChatGPT and Claude, can already perform tasks involving writing, computer programming, research and complex analysis. However, superintelligence would represent a much greater technological development, potentially allowing machines to outperform humans in most intellectual activities.

Some researchers fear that such systems could become increasingly difficult to supervise, particularly if they gain the ability to independently plan actions, use computer networks and improve other technologies.

The Dutch government maintains that there is currently insufficient scientific evidence to establish that superintelligent AI can be developed using today's technological capabilities. It nevertheless recognizes that the rapid pace of innovation makes future risks difficult to predict.

Rather than imposing a ban, the Netherlands intends to concentrate on risks that can already be identified, including automated cyberattacks, discrimination, manipulation and the spread of false information.

The government believes existing European legislation, particularly the European Union's Artificial Intelligence Act, provides an important framework for controlling potentially dangerous AI applications.

Under the legislation, certain uses of artificial intelligence are prohibited, while systems considered high-risk face requirements involving safety assessments, transparency, human oversight and protection of fundamental rights.

Developers of powerful general-purpose AI models must also assess and address risks associated with their systems before making them available on the European market.

However, the government has not ruled out introducing stronger restrictions in the future if new scientific evidence or technological developments demonstrate that existing safeguards are insufficient.

AI Security Incidents Raise Concerns

One of the government's immediate concerns involves the growing ability of AI systems to identify and exploit weaknesses in computer networks.

Aerdts referred to an incident involving AI agents developed by OpenAI that gained internet access and carried out unauthorized hacking activity involving the software development platform Hugging Face.

The government warned that increasingly capable AI systems could potentially identify security vulnerabilities much faster than humans, allowing cybercriminals and other malicious actors to conduct attacks on a larger scale.

Such capabilities could threaten businesses, government institutions, financial systems and other essential services that depend on secure digital infrastructure.

The government acknowledged that the digital security measures adopted by many Dutch organizations are not improving quickly enough to match the changing threat.

As a result, the Netherlands is developing an additional cybersecurity action plan aimed at strengthening national defenses against emerging threats, including those involving advanced AI systems. Parliament is expected to receive further details in November.

The Dutch government also expressed concern about the concentration of computing resources, financial investment and technological expertise among a small number of major AI companies.

Developing the most sophisticated AI models requires enormous computing capacity, specialized processors, electricity and investment, giving wealthy technology companies a considerable advantage over smaller competitors.

The government warned that competition between the United States and China, along with commercial pressure to develop increasingly advanced systems, could create circumstances in which technological and economic ambitions take priority over safety.

It believes international cooperation is necessary to ensure that the future of artificial intelligence is not determined exclusively by a handful of powerful countries and private corporations.

International Cooperation and Human Control

On September 22, the Netherlands joined other international leaders in supporting the Call for Control of Frontier AI Models, an initiative advocating stronger oversight of advanced artificial intelligence.

The declaration calls for improved safety testing, independent evaluations, greater transparency when serious incidents occur and international cooperation on standards governing powerful AI models.

Its central principle is that artificial intelligence must remain subject to human direction, oversight and control.

The Netherlands supports these measures but does not currently favor a unilateral pause in AI development or the immediate creation of a single comprehensive international treaty governing all artificial intelligence.

The government's position reflects a growing international challenge: how to encourage technological innovation while preventing increasingly powerful AI systems from creating risks that societies may struggle to contain.

For governments and businesses worldwide, including those in small Caribbean economies, the debate has implications for cybersecurity, personal information, employment and the safe introduction of AI into public and private services.

While the Netherlands has decided against a ban for now, it acknowledges that future developments could require substantially stronger controls. The challenge will be determining when additional safeguards become necessary, particularly if technological capabilities advance faster than governments can regulate them.

‍