Anthropic says it disrupted an Iran-linked user that used its Claude artificial intelligence models to track U.S. naval forces, compile targeting material and research vulnerabilities in shipboard ...