新闻标题：OpenAI发布AI安全指南，强调董事会有权阻止新AI模型发布
OpenAI，全球领先的人工智能研究机构，近日发布了一份名为“准备框架”（Preparedness Framework）的AI安全指南。该指南旨在为公司内部的AI模型开发和维护提供一套操作规程，以确保AI的安全性和可控性。
值得注意的是，尽管这套指南仍处于测试阶段，但它明确规定了一个重要的原则：即使公司的CEO或其他领导层认为AI模型是安全的，董事会仍有权阻止其发布。这一规定体现了OpenAI对AI安全问题的高度重视，也表明了其在确保AI模型安全性方面的坚定决心。
OpenAI表示，这一规定的出台是为了在AI模型的开发过程中，确保所有的决策都经过了严格的安全评估。即使在最乐观的情况下，也不能保证AI模型的所有可能风险都能被完全消除。因此，董事会的介入可以作为一种有效的风险管理措施，防止可能存在的安全风险被忽视或低估。
此外，OpenAI还强调，这份指南并非是对公司内部决策的干预，而是一种责任和义务的表现。通过这种方式，OpenAI希望能够进一步提高AI模型的安全性，保护用户的利益，同时也为AI技术的发展提供一个更加安全、可控的环境。
英语如下：
News Title: OpenAI Releases AI Safety Guideline: Board of Directors Has the Right to Block the Release of New Models
Keywords: OpenAI, safety guideline, board of directors
News Content: News Title: OpenAI Releases AI Safety Guideline, Stressing that the Board of Directors has the Right to Block the Release of New AI Models
OpenAI, the world’s leading artificial intelligence research institution, has recently released an AI safety guideline called the “Preparedness Framework”. The guide aims to provide a set of operating procedures for the development and maintenance of AI models within the company, ensuring the safety and controllability of AI.
It is noteworthy that, although this guide is still in testing phase, it clearly stipulates an important principle: even if the CEO or other leadership believes that the AI model is safe, the board of directors still has the right to block its release. This provision reflects OpenAI’s high attention to AI security issues and also demonstrates its firm determination to ensure the safety of AI models.
OpenAI stated that the introduction of this regulation is to ensure that all decisions in the development process of AI models have undergone rigorous security evaluations. Even in the most optimistic scenario, it cannot guarantee that all possible risks of AI models can be completely eliminated. Therefore, the intervention of the board of directors can serve as an effective risk management measure to prevent potential security risks from being ignored or underestimated.
Furthermore, OpenAI emphasized that this guideline is not an interference with internal company decisions, but a manifestation of responsibility and obligation. In this way, OpenAI hopes to further enhance the safety of AI models, protect users’ interests, and provide a safer and more controllable environment for the development of AI technology.
【来源】https://www.cls.cn/detail/1547583
Views: 1