According to foreign media reports, the US government is considering a series of new AI control strategies aimed at restricting the export of specific AI models, particularly targeting countries such as Russia, China, North Korea, and Iran. This decision will specifically focus on the core software of advanced AI models such as ChatGPT, implementing strict export controls. This new strategy is seen as an extension and upgrade of the US government's control policies in the high-tech field over the past two years. Previously, the US government has taken a series of measures to restrict the export of high-end AI chips to China, aiming to slow down China's development in this field. Now, the US government is further expanding the scope of control to AI models themselves. Currently, the US government has taken action to prevent AI chips and their manufacturing equipment from flowing into China, and requires AI companies to report to the government when foreign customers use their services to train AI models that could be used for cyber attacks. However, there are currently no specific regulatory measures for the export of AI models. Large AI companies in the United States, such as OpenAI, Google DeepMind, and Anthropic, have all developed leading AI models. In the absence of effective government regulation, these models theoretically can be sold to individuals or entities worldwide. The US government is concerned that these advanced AI models may be used by competitors for cyber attacks against the United States or for the creation of biological weapons. Therefore, the United States is planning to take new control measures to ensure the secure use of these sensitive technologies. According to insiders, the United States may formulate export control measures based on the computational threshold set in the AI executive order issued in October last year. This means that any AI model that meets this threshold needs to report its development plans and test results to the Department of Commerce. However, according to the latest data from Epoch AI, there are currently no AI models that meet this computational threshold. However, analysts point out that this does not mean that the US export control threshold will remain unchanged forever. On the contrary, the United States may formulate more detailed and stringent control requirements based on actual situations and technological developments, such as stricter scrutiny of data types or potential uses of AI models. This series of control measures will undoubtedly have a profound impact on the development and application of AI technology worldwide.