How Anthropic has doubled down on AI safety

Anthropic, which was formed by a group of safety-conscious defectors from OpenAI, spent 2023 establishing itself as one of just a handful of envelope-pushing developers of commercial AI models.

The company raised almost $7 billion in 2023 (from Amazon and Google, among others), and last October it told investors that it was generating revenue at a $100 million annualized run rate. In July, the startup launched the second major release of its LLM, dubbed Claude 2, with improved performance on standardized tests, a larger memory for text input, and improved computer coding ability.

Numerous AI app developers use Claude 2’s API (often alongside those of other state-of-the-art models) for its unique skills in text summary and generation. On the safety side, the company implemented a novel way of keeping Claude away from toxic or dangerous content—an approach the company calls “Constitutional AI.”

Instead of laboriously providing a model with human feedback, the researchers give a model a list of general principles (a “constitution”) to follow when creating its responses. A second AI then monitors the first, continually critiquing how well it follows the principles.

Explore the full 2024 list of Fast Company’s Most Innovative Companies, 606 organizations that are reshaping industries and culture. We’ve selected the firms making the biggest impact across 58 categories, including advertising, artificial intelligence, design, sustainability, and more.

https://www.fastcompany.com/91033423/anthropic-most-innovative-companies-2024?partner=rss&utm_source=rss&utm_medium=feed&utm_campaign=rss+fastcompany&utm_content=rss