Anthropic Releases Advanced AI Model
Anthropic released its Claude Fable 5 AI model to the public, acknowledging risks despite implementing safeguards, following earlier concerns about its power.
Atlas Newsdesk ·
Anthropic Releases Advanced AI Model
Anthropic, an artificial intelligence company, released its Claude Fable 5 AI model to the public on Tuesday, May 28, 2024, despite prior internal concerns regarding its power and potential risks. This public release follows an earlier private preview of a related model, Claude Mythos, which raised apprehension among technology, finance, and government leaders due to its advanced capabilities.
The company stated that Claude Fable 5, a version of Claude Mythos, includes safeguards and user limitations. However, Anthropic acknowledged that "releasing a model this capable comes with risks" and that "Fable's capabilities exceed those of any model we've ever made generally available." The initial private release of Claude Mythos in April was restricted to a small group of organizations due to its perceived intelligence and potential for misuse.
Concurrently, Anthropic granted approximately 150 organizations, previously involved in the Mythos preview, access to Claude Mythos 5. This version reportedly lacks limitations on cybersecurity or biology for specific organizational uses. Anthropic indicated this access is currently limited to "cyberdefenders and infrastructure providers" but plans to expand it through a broader trusted access program. Both Fable and Mythos models are designed to operate "unattended" on human commands for extended periods, surpassing previous Claude iterations.