Description
Anthropic is an AI safety company founded in 2021, creator of the Claude model family. Known for research contributions to AI alignment, interpretability, and responsible scaling. Produces the Constitutional AI (CAI) alignment methodology and Claude (all versions).
Key Properties
- Founded: 2021
- Location: San Francisco, CA
- Founders: Dario Amodei, Daniela Amodei, and others from OpenAI
- Focus: AI safety research, frontier model development, responsible scaling
- Notable models: Claude (1, 2, 3, 3.5, and Sonnet/Opus/Haiku variants)
Notable Sources
- constitutional-ai-harmlessness-from-ai-feedback — Constitutional AI (2022); the training methodology that underpins Claude’s harmlessness
Related Entities
- openai — Anthropic’s founders came from OpenAI