Why AI systems may never be secure, and what to do about it

You might, for example, innocently instruct an AI agent to summarise a thousand-page external document, cross-reference its ...