OpenAI updated itsArtificial Intelligence Safety Principles, which states that even if a company's leadership believes its artificial intelligence technology is sufficiently secure, the board of directors still has sufficient authority to restrict its implementation plans.
In the new security principles, artificial intelligence technology, which is promoted by senior executives including current CEO Sam Altman, will have the right to restrict or overturn technology implementation plans as long as the board of directors believes there are potential risks.
This approach highlights that the board still has sufficient authority to regulate the development of artificial intelligence technology promoted by OpenAI to avoid potential safety risks and catastrophic consequences, such as causing hundreds of billions of dollars in economic losses or harming people's lives.
Currently, OpenAI oversees the development of its AI technology through three independent teams: "Preparedness," "Safety Systems," and "Superalignment." These teams assess whether AI poses a threat in different areas, identify security vulnerabilities in the technology, and guard against the emergence of uncontrollable AI technology in the future.
OpenAI also calls on other companies to use the same safety principles to assess the potential risks of AI, categorizing AI as low, medium, high, and critical risk, and to determine the practical benefits of AI applications. Under the new safety principles, OpenAI stated that it will only release AI application models with low and medium risk assessments.
