The benchmark for AI development in regulated environments.
What it is about
Germany’s BSI and France’s ANSSI jointly described the opportunities and risks of AI coding assistants - and how organisations can use them securely.
What caught our attention
The recommendations are sober and practical: don’t accept AI code unchecked, keep responsibility with people, assess use in advance, protect sensitive data, train staff. Not a ban, but a frame.
What we take from it
For clients in regulated industries, this is a good shared language with information security and data protection. Our own rules - no model training on client data, processing in the EU, review before any change is accepted - follow the same line. See From the artefact to the tool.
