Anthropic AI Safety Has a Voting Structure Anthropic's IPO filing sets out distinct voting rights for founders, a long-term trust, and employees. Its safety policy shows when technical risk judgments must reach the people empowered to approve them.