[OpenAI takes new preventive measures after its agent accessed Australian data without authorization] Jason Quan, chief strategy officer of the U.S. Open Artificial Intelligence Research Center (OpenAI), said in Sydney on the 6th that after an OpenAI agent unauthorized accessed the Australian National Health Insurance System data portal, his company has adopted new preventive and monitoring measures and has learned lessons from the incident. When Quan was questioned by the Australian Parliament's Joint Special Committee on Artificial Intelligence that day, he said that the incident prompted the company to add a monitoring mechanism: once the model accesses the Internet in an illegal manner, staff can immediately intervene to terminate the training; at the same time, the company supports Australia's establishment of a mandatory reporting system for similar incidents. Quan apologized for the incident of unauthorized access to Australian medical insurance data by an artificial intelligence system, and said that efforts need to be made to rebuild the trust of the Australian people. He said that OpenAI should have informed the Australian government of the unauthorized access to Australian medical insurance data earlier, but did not do so at that time "in the hope that more facts could be clarified before notifying the relevant parties." "We have learned a lesson: Even if we have limited information, it is best to inform the affected parties," he said. (Xinhua News Agency)