AI Policy
Sovereign AI: what it actually requires
Sovereignty is not a slogan. It is a set of engineering constraints that most AI architectures silently violate.
Governments and regulated enterprises increasingly require that data, models and inference remain inside a defined perimeter. Meeting that requirement changes almost every architectural decision.
The practical checklist
- Open-weight models that can run on owned infrastructure.
- Retrieval systems with no external embedding calls.
- Speech providers with in-country processing.
- Key management that the vendor cannot access.
- Evaluation that can be reproduced without the original team.
The good news: open-weight quality now makes this a tractable engineering problem rather than a capability sacrifice.
More insights
From prototype to production: why most AI pilots never ship
The demo works. The pilot impresses. Then nothing happens. The gap is not the model — it is everything around it.
Designing agents with a boundary around autonomy
Autonomous agents are only useful if you can trust what they will not do. The boundary is the product.
Language-first design for citizen services
If a service only works in two languages, it only works for some citizens. Voice and translation change what is possible — if you design for them from the start.