Anthropic

Anthropic. It is an American artificial intelligence (AI) startup company and public-benefit corporation founded by former members of OpenAI. The company focuses on developing reliable, interpretable, and steerable AI systems, with its flagship product being the Claude family of large language models. Anthropic's research emphasizes AI safety and AI alignment, seeking to ensure advanced AI systems behave in accordance with human intentions and values.

History

The company was established in 2021 by siblings Dario Amodei, previously VP of Research at OpenAI, and Daniela Amodei, who held roles at OpenAI and Stripe. Their departure was influenced by disagreements over the direction and safety priorities of OpenAI. Key early team members included several other former OpenAI researchers. In its initial phase, Anthropic secured significant early funding from investors like Jaan Tallinn, the co-founder of Skype, and the Center for Emerging Risk Research. A major milestone was reached in 2023 with a multi-billion dollar investment from Amazon, followed by a further substantial commitment from Google via its Google Cloud platform. These partnerships integrated its models into services like Amazon Bedrock and positioned it as a leading entity in the competitive Generative artificial intelligence landscape alongside firms like OpenAI and Google DeepMind.

Products and services

The primary offering is Claude, a conversational AI assistant. The model has been released in successive iterations, including Claude 2 and Claude 3, which are accessible through a public-facing chatbot interface and an API for developers. Claude is integrated into productivity platforms such as Slack and is available to subscribers of Quora's Poe. The company also provides Claude for Business, a suite with enhanced administrative controls and security features tailored for enterprise clients. Its models are deployed as a service on major cloud infrastructures, notably Amazon Web Services and the Google Cloud Platform, enabling broad commercial and research application.

Research and development

Core research is centered on Constitutional AI, a novel training framework designed to align AI systems with a set of written principles, or a "constitution," thereby reducing reliance on extensive human feedback. This work is detailed in influential papers such as "Constitutional AI: Harmlessness from AI Feedback." The research team, which includes noted scientists like Chris Olah, focuses on interpretability, attempting to reverse-engineer the internal representations of neural networks. This field, sometimes called mechanistic interpretability, aims to make model decision-making processes more transparent. The company actively publishes its safety research and collaborates with academic institutions and policy organizations like the Centre for the Governance of AI to advance the broader field of AI safety.

Corporate affairs

Anthropic is structured as a Delaware public-benefit corporation and a Delaware certified B Corporation, legally obligating it to balance public good with profit. Its corporate governance includes a Long-Term Benefit Trust designed to steward the company's mission over extended time horizons. The leadership team is led by Dario Amodei as CEO and Daniela Amodei as President. Major financial backers, beyond Amazon and Google, include Zoom Video Communications, Salesforce, and Menlo Ventures. The company has engaged with U.S. policymakers, providing testimony before the United States Senate Judiciary Subcommittee on Privacy, Technology and the Law regarding the regulation of foundation models.

See also

* OpenAI * Google DeepMind * Large language model * AI alignment * Dario Amodei

Category:Artificial intelligence companies Category:Companies based in San Francisco Category:2021 establishments in California