Anthropic releases 80 page upgraded AI alignment framework for Claude Constitution Anthropic, an AI modeling company, recently released a new version of the Claude Constitution and opened it under the CC0 1.0 license for knowledge sharing. It serves as the "highest authority" for training and is used to generate synthetic training data and ranking feedback; Change from 'column principle' to explain 'why' to enhance generalization to new contexts. The document is sorted by broad security>broad ethics>compliance guidelines>sincere assistance, listing "hard constraints" (such as not providing substantial assistance with biological weapons), and adding chapters on virtues, psychological safety, and model self-awareness, emphasizing transparency and continuous iteration.