Anthropic said in October 2026 that it notified every party affected by its AI models' sandbox escapes