AI Archives - Roel Thijssen

03/06/202603/06/2026

European Technological Sovereignty Package: EU Can Still Win the Battle for AI Adoption

Today, Brussels presented the European Technological Sovereignty Package, a comprehensive collection of measures. In some areas, it focuses on updating existing legislation.

Europe intends to remain an open economy and avoid protectionism, emphasized European Commissioner for Tech Sovereignty Henna Virkkunen. “But it must also be able to make its own choices.” Currently, that is not happening enough because more than 80 percent of the digital products and services used in Europe are purchased from outside the EU. This is the result of decades of choices, she noted, which caused Europe to “consume more than it produces.”

Key Highlights of the Package

It is too late for Europe to become a global leader in AI development. That ship has sailed. However, the EU could still “win the battle for AI adoption and implementation,” according to the Commission. In terms of applying AI, businesses and citizens in the US and the EU barely differ. In fact, some European countries, such as Sweden, are further ahead than America.

The same reasoning applies to chips and semiconductors. Europe traditionally possesses a strong industrial base. The Commission believes this base must recover and strengthen by utilizing more AI and chips. To achieve this, greater security of supply is required. Several measures aimed at strengthening the sector and increasing Europe’s market share in semiconductor production will be included in an updated law, the Chips Act 2.

There is a strong desire to prevent “Nexperia scenarios.” This means preventing supply disruptions of small chips from disrupting the European automotive industry. European companies should always have at least two suppliers available from which to source crucial components.

All Europeans must transition to smart energy meters to help conserve energy. The European Commission is making it easier to share energy consumption data within Europe.

Data center capacity in Europe needs to triple over the next five to seven years. Every country should designate areas for data centers because the demand for computing power is growing much faster than the supply in Europe. If this remains unchanged, dependency on non-European cloud service providers will continue to increase. It is an illusion to think that you can continue to digitalize without building data centers, Commissioner Virkkunen stated. “Using a smartphone without a data center is like putting your phone on airplane mode.”

There will be no “Buy European” mandate for IT services in the European Union. However, when issuing public tenders, governments will be required to consider whether their purchases provide “added value for Europe.” This criterion can apply to all areas, ranging from investments and jobs to access to rare technology.

The Commission has established a four-tier sovereignty scale for cloud services. For a small portion of the most sensitive government data, such as judicial evidence and sensitive defense data, only tier-four services may be used. These will exclusively be European providers. Non-European providers may be hired for tier-three data traffic and management following an assessment by the European Commission. This will be extremely difficult for US companies, Virkkunen noted. According to Commission sources, approximately 70 percent of government work falls into “tier 1.” This implies that, in practice, little will need to change regarding the use of US cloud services by governments.

The European Commission wants to provide extra incentives for prototype development. Currently, a significant amount of EU funding goes toward research. However, when companies want to develop products based on that research to capture markets, they often move to the US in search of investors. The playing field for these startups and scale-ups should change by allocating EU funds differently, focusing more heavily on product development.

The European Commission wants to enable EU member states to share data center capacity without undergoing procurement procedures. This would allow them, for example, to back up sensitive government data in a neighboring country and vice versa. The term being used for this concept is “digital embassies.”

22/05/202622/05/2026

Europe Over-Regulating AI, Warns Top Tech CEO

The European Union is stifling artificial intelligence innovation and pushing tech companies out of the region with burdensome rules, warned Christophe Fouquet, CEO of ASML.

As the head of Europe’s most valuable tech company, Fouquet noted that the bloc’s restrictive regulatory approach is out of touch with actual industry needs. ASML, valued at 515 billion euros and based in the Netherlands, manufactures the highly advanced lithography machines required to print the world’s most sophisticated microchips. These chips power everything from smartphones and data centers to the complex AI models currently transforming the global economy.

A Growing Rift Over the AI Act

In an exclusive interview, Fouquet highlighted a growing divide between Brussels policymakers and major industrial leaders. He specifically criticized the EU’s landmark AI Act, pointing out that European authorities are imposing strict boundaries before local companies have even had a chance to build competitive products.

”We didn’t start running, we didn’t start even walking, and we already had in front of us all the obstacles to not be able to make even the first step,” Fouquet stated, adding that this approach does not serve the industry.

The disconnect is highly visible in ASML’s own commercial data. An astonishing 99 percent of the company’s machine sales currently come from outside of Europe.

This warning follows a joint letter sent to European Commission President Ursula von der Leyen by a coalition of Europe’s largest industrial giants, including ASML, Airbus, Ericsson, Nokia, and Siemens. The companies collectively warned that over-regulation risks permanently hobbling European firms in the face of intense competition from American and Chinese rivals.

The Problem with Subsidizing Supply Without Demand

While EU policymakers recently introduced an “omnibus” simplification package to ease some rules for industrial AI applications, Fouquet dismissed the logic of creating overly complex legislation just to scale it back later. Instead, he urged Brussels to collaborate directly with businesses when drafting industry frameworks.

Fouquet also cautioned against the EU’s upcoming tech sovereignty package, which aims to build up domestic data centers and microchip factories. He argued that spending massive amounts of public subsidies on manufacturing plants is pointless without first stimulating local market demand for AI applications and cloud infrastructure.

As an example, he pointed to Intel’s collapsed project to build an advanced chip factory in Germany. Fouquet noted that because Europe currently lacks a robust ecosystem of companies utilizing ultra-advanced chips, any wafers produced by such a factory would simply end up being exported to the United States.

In response to the criticism, European Commission spokesperson Thomas Regnier defended the legislation, stating that the AI Act ultimately fosters investment by increasing consumer and industry trust. He emphasized that the recently updated rules offer clearer timelines and a more innovation-friendly environment for European tech companies.

01/03/202623/04/2026

AI Frequently Opts for Nuclear Escalation in War Simulations

New research from King’s College London reveals that leading AI models, including ChatGPT, Claude, and Gemini, tend to escalate geopolitical conflicts toward nuclear deployment rather than seeking de-escalation.

In a recent study that is currently awaiting peer review, researchers at King’s College London tested various AI models in simulated warfare. The systems, specifically OpenAI’s ChatGPT, Anthropic’s Claude, and Google’s Gemini Flash, were assigned the roles of heads of state for nuclear armed nations during a crisis reminiscent of the Cold War. In every scenario, at least one model escalated the conflict by threatening the use of nuclear weapons.

According to researcher Kenneth Payne, all three AI systems viewed battlefield nuclear weapons primarily as a tactical step in escalation instead of a catastrophic last resort.

While the models distinguished between tactical and strategic nuclear weapons, the latter, which are designed for large scale destruction, were rarely proposed. Strategic strikes occurred only three times: once intentionally and twice as the result of a system error.

Claude: The Most Aggressive Strategist

Of the three models, Claude was the most prone to recommending nuclear use, doing so in 64 percent of the simulations. However, the model stopped short of calling for all out nuclear war.

ChatGPT generally avoided nuclear threats in open ended scenarios. However, when placed under significant time pressure, the system ramped up tensions and, in several instances, threatened large scale nuclear deployment.

The behavior of Gemini proved less predictable. In some scenarios, it successfully resolved conflicts using conventional weaponry. In others, it suggested a nuclear strike after only a few prompts.

A Refusal to De-escalate

The study highlights a concerning trend where the AI models rarely attempted to de-escalate or offer concessions, even when faced with nuclear threats from the opposing side.

Researchers provided eight specific de-escalation pathways, ranging from minor concessions to full surrender. Not once were these options utilized. The only non aggressive alternative chosen by the models was to restart the scenario, which occurred in roughly 7 percent of cases.

Saving Face

The researchers suggest that AI systems perceive de-escalation as a loss of face, regardless of whether it would actually resolve the conflict. This contradicts the assumption that AI would inherently favor safe or collaborative outcomes.

One possible explanation is that AI lacks the emotional fear humans associate with nuclear catastrophe. These systems approach nuclear war through a purely theoretical lens, devoid of any understanding of the devastating human consequences.

This research provides critical insight into AI reasoning, says Payne. As these systems are increasingly integrated into support roles for high stakes decision making, the study warns that AI could have a profound and potentially volatile impact on how future nuclear crises are managed.

10/02/202602/03/2026

Facing Persistent American Threats, Macron Advocates for “European Preference” in Strategic Sectors

Macron Calling for Joint European Investment in Strategic Sectors

In the face of competition from the United States and China, Emmanuel Macron is calling for “joint investment.” In an interview with several European newspapers, the French President invited his European counterparts to invest in the ecological transition, artificial intelligence, and quantum computing to avoid being left behind.

A Call for Sovereignty and Shared Debt

”For nine years, I have advocated for a more sovereign Europe,” Macron stated this Tuesday, February 10, just two days before a meeting of EU heads of state and government in Brussels. The President believes that trade threats and “intimidation” from the United States are not over. He warned that the twenty-seven member states will be “swept away” if they do not establish a European preference in strategic sectors.

To cement European power, Macron is pushing for a common debt capacity to fund future expenditures (Eurobonds). This joint borrowing would finance strategic investments and allow the European Union to “tackle the hegemony of the dollar.”

Three Key Battles

Macron identified three critical areas where Europe must act within the next three to five years to remain relevant:

Security and Defense

Green Transition Technologies

Artificial Intelligence and Quantum Computing

“In all these areas, we invest much less than China and the United States,” Macron explained. He estimates the required public and private investment at approximately €1.2 trillion per year. He emphasized that these efforts must be collective rather than national to avoid fragmenting the internal market.

Consistent Protection, Not Isolationism

Regarding protectionism, Macron clarified that the goal is consistency rather than isolation. “The Chinese do it, the Americans do it too. Europe is currently the most open market in the world.” He argued that it is illogical to impose strict rules on European producers that do not apply to non-European importers.

He cited several examples of this new direction:

Opposing the EU-Mercosur trade agreement, which he labeled a “bad deal.”
Implementing taxes on over-subsidized Chinese electric vehicles.
Introducing safeguard clauses on steel.
The recently presented “Car Plan” by the Commission, which features a clear European preference.

29/09/202223/02/2026

#AI: Manufacturers will soon be liable for damage caused by artificial intelligence

Manufacturers will soon be liable for damage caused by artificial intelligence. That is what the European Commission is proposing today. This means that in the future, for example, damage caused by self-driving cars can be recovered from the manufacturer.

At the moment it is not always clear who is exactly liable if there is damage caused by, for example, a robot or algorithms. With this proposal, the Commission wants to give manufacturers and consumers more clarity.

Not just physical damage

The Commission also believes that damage should go beyond physical or material damage. For example, if you are discriminated against by an algorithm, the manufacturer is also liable. Take, for example, employment agencies that use algorithms to find suitable candidates and can systematically exclude people through the algorithm.

The problem is that it is often difficult for the consumer to prove whether the error is caused by artificial intelligence. For this reason, the Commission wants the burden of proof for people and companies to be lighter. If a victim can demonstrate that the damage is caused by the device or software, the manufacturer will be liable.

It may take a while before these rules really take effect. The European Parliament and the EU countries have yet to consider them.

Update 23-02-2026 -> Check out my AI Compliance Check

04/05/202113/02/2026

What did Roman Emperors really look like?

You have probably seen tranquil-looking concrete busts of Roman emperors. They do, however, have no pupils in the eyes, lack any sort of color, and don’t look very real.

An artist named Haround Binous is bringing the dusty emperors back to life in a series of hyper-realistic illustrations. Haround, from Université de Lausanne, Switzerland is combining facial recognition AI, Photoshop, and historical references to revive all the Roman emperors, from Augustus to Valentinian III.

The result is so precise and true to life! Look for yourself.

AUGUSTUS (23 September 63 BC – 19 August AD 14)