As artificial intelligence (AI) systems are increasingly deployed in high-stakes domains like national security and warfare, the challenge of aligning these technologies with human values takes on new urgency and complexity. But in contexts where harm is an intended outcome, what does “value alignment” really mean?
