Anthropic weighs new AI model despite CEO's safety warnings

Anthropic, the AI company led by Dario Amodei, is reportedly considering launching a new model to compete with OpenAI, just days after Amodei publicly urged the industry to slow down AI development for safety reasons. This potential move highlights a tension between safety concerns and market competition. According to Reuters, citing three sources, Anthropic is evaluating a response to OpenAI's GPT-6 Astra, released on September 3, which has already captured about 13% of AI spending on the Ramp platform versus 8% for Anthropic's Claude Fable. The new model has no release date yet, and the company is still assessing its safety risks. Anthropic is also preparing for a possible stock market listing after the November US elections, with annualized revenues exceeding $65 billion. Experts note that in this competitive landscape, no company wants to be the first to slow down, and Amodei's safety warnings may also serve to shape future regulations and reassure investors.