Anthropic’s latest disclosure reveals that its AI model Claude is being used to help build unspecified systems, offering rare insight into a frontier AI lab. The disclosed figures represent an early effort to measure a difficult question about the model’s impact.