Anthropic is an AI safety and research company that develops the Claude series of large language models, focusing on building AI systems that are steerable, interpretable, and safe.