Anthropic hires its first "AI welfare" researcherCompanies should assess AI consciousness using multiple speculative indicators, balancing moral consideration while avoiding the pitfalls of anthropomorphism.