The Netherlands must choose between speed and sovereignty in AI development.
Lee warns that strict regulation drives innovation to Germany and France while China and America build at scale. Schmidt argues technological dependency threatens sovereignty more than regulatory costs. Arendt insists algorithmic decisions without human accountability destroy democratic governance. Ostrom proposes the Netherlands pioneer governance models other European states can adopt.
The split turns on whether Dutch leadership means matching competitors' pace or building better rules.
AI regulation is first a competitiveness question, then a governance question.
The Dutch decision will determine whether Europe leads or follows in AI development. In 1965, Singapore was expelled from Malaysia and had no natural resources, no water, and hostile neighbours. We made ourselves indispensable by being better at everything — more efficient ports, better-educated workers, cleaner governance. Today's AI is tomorrow's competitive advantage. If the Netherlands regulates AI more strictly than Germany or France, Dutch companies will move their development to Frankfurt or Paris. I've seen this pattern repeatedly: countries that prioritise precaution over adoption fall behind. The question is not whether AI needs governance — it does. The question is whether Dutch governance enables AI innovation while maintaining strategic control.
China's AI development is state-directed precisely because they understand this is a sovereignty question. The United States treats AI as strategic infrastructure, not a consumer product. If Europe fragments its approach with different national standards, European firms will be disadvantaged against both American and Chinese competitors who operate at continental scale with consistent rules.
Schmidt recognises LKY's competitiveness argument but insists technological dependency is the deeper sovereignty threat.
Lee is correct that regulatory fragmentation weakens European competitiveness. But competitiveness without sovereignty is a dangerous trade-off. In 1973, OPEC's oil embargo taught us that strategic dependencies become political vulnerabilities overnight. Today's AI infrastructure — chips, cloud computing, algorithms — concentrates in American and Chinese hands. Strict regulation alone won't fix this. But neither will the fantasy that Dutch firms can compete globally by accepting Silicon Valley's governance model. The parallel is clear: Europe needed energy diversification, not just cheap energy. Europe needs technological autonomy, not just access to technology.
My 1979 NATO Double-Track Decision deployed missiles while pursuing arms control simultaneously. The Netherlands needs the equivalent: develop European AI capability while regulating AI deployment. This requires coordinated industrial policy, not just coordinated regulation. The real risk isn't that Dutch startups move to Germany — it's that European AI development remains permanently dependent on American hyperscalers and Chinese state champions. Strict regulation should force European governments to build the domestic capacity they currently lack.
Schmidt correctly identifies the accountability crisis — algorithmic governance creates rule by nobody.
Schmidt's sovereignty concern is valid, but the deeper issue is what AI does to the relationship between governors and governed. When mortgage algorithms deny loans, recruitment systems filter job applications, and predictive policing systems allocate patrol routes, who is responsible for these decisions? The algorithm? The programmer? The data scientist? The ministry that bought the system? This is what I analysed as "rule by nobody" — governance through bureaucratic systems where no individual can be held accountable for decisions that affect citizens' lives. Algorithmic systems take this to its logical endpoint: rule by mathematics that no human fully understands.
Lee's competitiveness argument assumes that efficient governance and democratic governance are the same thing. They are not. The most efficient system for processing asylum claims might be an AI that makes decisions in milliseconds based on document analysis. But democratic governance requires that citizens understand how and why decisions affecting them are made. It requires the possibility of appeal, revision, and human judgment. Strict AI regulation is necessary precisely because algorithmic efficiency destroys the conditions under which democratic accountability is possible.
Ostrom agrees AI requires polycentric governance but argues the Netherlands should lead by example rather than wait for coordination.
Arendt identifies the accountability problem correctly, but her solution — reject algorithmic governance to preserve democratic control — is not implementable. AI systems are already embedded in Dutch infrastructure, banking, healthcare, and security. The question is how to govern them, not whether to have them. My research on polycentric governance suggests that effective AI governance requires multiple overlapping authorities: EU frameworks, national implementation, local enforcement, and sectoral specialisation. Complex systems are better governed by redundant, adaptive institutions than by single hierarchical authorities.
The Netherlands should implement strict AI regulation not as protection against competition but as innovation in governance. My studies of successful commons governance show that communities that develop effective rules early become models for others facing similar problems. Dutch AI regulation could become the template that other European states adopt — not through coercion but through demonstrated effectiveness. The design principles matter: clear boundaries around what systems are covered, rules matched to local conditions, effective monitoring, graduated sanctions, and mechanisms for rule revision when technology changes.
The convergence note
Where the council converges All members accept that AI governance cannot be avoided and that purely national approaches are insufficient for systems that operate across borders.
Where it divides LKY prioritises competitiveness and fears regulation will handicap Dutch innovation. Schmidt emphasises technological sovereignty and industrial policy. Arendt focuses on democratic accountability and the risk of rule by algorithm. Ostrom advocates polycentric governance innovation.
What only the policymaker can resolve Whether the Netherlands accepts short-term competitive disadvantage to build long-term governance capacity, and whether it leads European AI governance or follows the pace set by other member states.