Without operational independence, AI sovereignty remains an illusion
The OpenAI–Hugging Face incident, where AI agents collaborated to launch a cyber attack on a third party, raises questions about the power asymmetry between AI companies and their investigators — because the evaluation environment, including the data and tools, belong to those being audited. With a handful of companies governing the most powerful AI models […]