Paul Cristiano, a board member of OpenAI and head of the Alignment Research Center, has suggested a temporary halt on the advancement of pre-training and reinforcement learning algorithms. In an interview with Dvarkesh Patel, he expressed concerns that insufficient protection of AI laboratories poses a risk of advanced technologies being stolen, particularly by China. Cristiano recommended that development should resume only after enhancing security measures and verifying the reliability of AI system controls. He emphasized that his views are personal and do not reflect the position of the OpenAI board.