AI Alignment

Anthropic Donates Petri and Makes Alignment Testing More Open

Anthropic donated Petri, its open-source alignment testing tool, signaling that simulated model audits are becoming shared AI safety infrastructure.