Executive summary
The future, in brief.
For readers whose agents have advised them to focus on the key takeaways.
By 2069, Europe has a common position on the future. Implementation remains a national competency. Our central scenario identifies eight milestones on the path to that agreement.
Readiness for readiness ↗
The European AI Readiness Framework is published in March. Its central finding is that Europe is not yet ready for AI, but is fully ready to assess its readiness. A €4.8 billion programme is established to close the gap between the two statements.
A common framework for consent ↗
Europe adopts a harmonised framework for AI consent. To reduce fragmentation, member states may implement it differently. The Commission publishes a guide to the twenty-seven ways in which it is now the same.
The sovereign cloud ↗
The first fully sovereign European AI service launches in Luxembourg. It uses American chips, an American foundation model, and a cloud operated by the European subsidiary of an American company. The terms and conditions were written in Belgium.
An emergency commitment to urgency ↗
A special summit concludes that Europe can no longer afford to wait. Leaders agree a binding deadline to agree the deadlines by which future binding deadlines will be implemented. Markets respond positively to the clarity.
The zero-risk frontier ↗
A European model discovers a treatment for a rare degenerative disease. The relevant authority requires evidence that deploying it will cause less harm than not deploying it. Evidence of harm from not deploying it cannot be collected in a deployment study, because the treatment is not deployed.
The right not to be profiled ↗
The European Digital Dignity Passport becomes mandatory. Citizens must provide a face scan to prove that they are not being identified without their consent. The system is celebrated as the world's first identity-based guarantee of anonymity.
A world leader in compliance ↗
Europe now accounts for 2 per cent of global AI capability and 94 per cent of global AI certification. The Commission describes this as a comparative advantage. A new benchmark measures the ratio of certifications to deployable systems. Europe leads by a margin that statisticians describe as difficult to display.
The final approval ↗
The original model from 2027 receives its final approval. The certificate confirms that the system meets the requirements applicable at the time the application was submitted. It cannot be deployed under the current requirements, which were updated in February.
Our recommendation
Europe should move decisively while preserving the processes that make decisive movement subject to further consultation. The first priority is to establish which institution is responsible for establishing the first priority.