Germany's National Security Council greenights an AI Safety Institute modeled after the UK's AISI
Germany just created its own AI safety institute. But there's a catch: it has to rely on American and Chinese models to do it.
Why it matters
Germany's new DE-AISI represents a critical shift in how EU nations are approaching AI governance and security testing, but exposes a strategic vulnerability: Europe's dependence on US and Chinese frontier models undermines its ability to maintain sovereign AI oversight.
The key facts
5 to knowGermany's National Security Council approved establishment of DE-AISI (German AI Safety Institute)
Model follows UK's AISI framework for frontier model security testing
Will test models from Anthropic and OpenAI
Highlights EU's strategic dependency on US and Chinese AI providers
Raises geopolitical concerns about AI sovereignty and government linkages to home nations
Go to the source
The Decoderthe-decoder.com
Publisher excerpt: Germany's National Security Council has decided to establish an AI security institute. The "DE-AISI" will test frontier models from Anthropic or OpenAI for security risks, following the British model. But as long as the EU doesn't have its own frontier models, it remains dependent on US and Chinese…