Live activity
Articles and posts from Adrian Hollister across the blog, LinkedIn and Bluesky: on digital sovereignty, cyber security, data protection, and the cost of depending on US infrastructure.
A question worth carrying into your next architecture review: if your identity provider disappeared behind someone else's law tomorrow, how long before you could open your own front door?
View on LinkedIn → Related: The Last American Anchor: The Login →European AI depends on foreign governments. This paper asks what they have done with that power and what Europe should conclude. 43 pages of evidence, 69 sources, free under CC BY-SA.
View on LinkedIn → Related: Your AI Answers to a Foreign Government. Part 1: China vs America →Your AI answers to a foreign government... China vs America. Europe spent two years asking whether Chinese AI was safe. February 2026 answered a different question. The US government ordered every federal agency to stop using Anthropic, an American company, because Anthropic refused to strip two…
View on LinkedIn → Related: Your AI Answers to a Foreign Government. Part 1: China vs America →Most of what Europe calls AI strategy is a brochure. I have a four-question test that separates strategy from brochure, and I have named it because named tests get used: the buyability test. Can a European organisation with a sensitive workload buy the capability it needs, under law it answers to…
View on LinkedIn → Related: What a Credible European AI Layer Would Take →In May 2025 a US court ordered OpenAI to preserve every ChatGPT conversation it would otherwise have deleted, across every consumer tier, including conversations users had already deleted. EU users were told their GDPR erasure rights were suspended. Later that year the court ordered 20 million of…
View on LinkedIn → Related: Your AI Answers to a Foreign Government. Part 1: China vs America →Washington has cut off AI access three times in two years: to China in 2024, to an Indian company in 2025, and to its own frontier lab, Anthropic, in 2026, over a refusal to enable mass surveillance. Beijing has never needed to cut anyone off. China's National Intelligence Law obligates every…
View on LinkedIn → Related: Your AI Answers to a Foreign Government. Part 1: China vs America →Your AI answers to a foreign government: China vs America. Washington has switched AI access off three times in two years, most recently against Anthropic, an American company punished for refusing mass domestic surveillance and autonomous weapons. The CLOUD Act reaches your data wherever it is…
View on LinkedIn → Related: Your AI Answers to a Foreign Government. Part 1: China vs America →Keep one way into your own systems that depends on no cloud at all. Local accounts, sealed credentials, tested now and then. The day your identity provider is unreachable should be a bad day, not the end of the story.
View on Bluesky → Related: The Last American Anchor: The Login →Ask a team how often their reviewers overturn the AI's output. Perhaps zero disagreement is success. Mostly it means nobody is checking any more. Fluent text does this. It arrives reading as though somebody already thought it through, so nobody does. The reviewer starts as an editor and ends as a…
View on LinkedIn → Related: The Approve Button Is Doing a Lot of Work →An EU region in an American cloud changes the geography, not the jurisdiction. The operator is the operator, and the law follows the operator. Residency is what the brochure sells. Sovereignty is what the courtroom tests.
View on Bluesky → Related: Residency Is Not Sovereignty →The last American anchor. You can move every workload to a European cloud and still leave the most important system in America: the one that decides who gets in. This piece exists because Brian J. asked a really good question under my Airbus post: what did Airbus do about identity? His way of…
View on LinkedIn → Related: The Last American Anchor: The Login →