DeepSeek Gets an ‘F’ in Safety From Researchers by AJ Dellinger from Gizmodo on 2025-02-04 15:00 (#6V1RY) The model failed to block a single attack attempt.