← All statements
Plenary speech

Reinier VAN LANSCHOT

The newest AI model from Anthropic, Mythos, is so effective at finding weaknesses in software that they decided to only give access to a few companies from the US. In the wrong hands, it can be used to exploit digital vulnerabilities […] at a speed and skill never seen before. […] The next AI model might not be kept under lock and key. It might be pointed at us!

EU cybersecurity and preparedness in view of advanced AI systems (debate)official record ↗

Read the full statement ↓

Madam President, dear colleagues, dear Commissioner, the newest AI model from Anthropic, Mythos, is so effective at finding weaknesses in software that they decided to only give access to a few companies from the US. In the wrong hands, it can be used to exploit digital vulnerabilities, like attacking bank accounts or health records at a speed and skill never seen before. Another wake-up call for Europe!

Anyone who believes the best AI will always be available to us is making a dangerous assumption. As soon as AI models become strategic assets for cybersecurity, defence or economic power, access will become selective. If we do not have AI made in Europe, if we do not control our own digital infrastructure, we cannot control our own future.

The next AI model might not be kept under lock and key. It might be pointed at us! So I ask you, Commissioner, to act quickly and build EU-owned infrastructure, data centres, cloud and compute on European soil. Invest in research and development of purpose-built AI, cybersecurity and open source standards. If we act quickly, we will still have a future made in Europe.

Text as published by the European Parliament, in the Parliament's own English rendering of the original language.