在瑞士日内瓦举行的第27届联合国科技大会上，世界数字技术院（WDTA）正式发布了两项重大国际标准，分别为《生成式人工智能应用安全测试标准》和《大语言模型安全测试方法》。这些标准的发布标志着全球科技巨头在推动AI安全规范方面迈出了重要的一步。
据悉，这两项标准是由OpenAI、蚂蚁集团、科大讯飞、谷歌、微软、英伟达、百度和腾讯等全球领先的科技企业以及众多专家学者共同编制完成的。这一跨行业的合作旨在确保人工智能，尤其是生成式AI和大语言模型在快速发展的同时，能够遵循严格的安全和伦理准则，保障用户隐私和数据安全。
《生成式人工智能应用安全测试标准》旨在为AI系统的设计和开发提供一套全面的安全评估框架，防止潜在的滥用和误导。而《大语言模型安全测试方法》则聚焦于语言模型的训练和应用过程，以确保其在处理敏感信息和交互时的可靠性和公正性。
此次国际标准的制定反映了全球科技行业对AI安全的重视，也体现了各企业在推动技术创新与社会责任之间的平衡。随着AI技术的广泛应用，这些标准将为全球范围内的开发者、企业和政策制定者提供重要的参考依据，促进AI技术的健康发展和全球数字生态的构建。
英语如下：
News Title: “Global Tech Giants Unveil International Standards for Large Model Safety at UN Science Summit”
Keywords: Large Model Safety, International Standards, Science Summit
News Content: At the 27th United Nations Science Conference in Geneva, Switzerland, the World Digital Technology Academy (WDTA) officially released two significant international standards: the “Safety Testing Standard for Generative Artificial Intelligence Applications” and the “Safety Testing Method for Large Language Models.” These launches signify a crucial step by global tech giants in advancing AI safety regulations.
It is reported that these standards were jointly developed by leading tech companies such as OpenAI, Ant Group, iFlytek, Google, Microsoft, NVIDIA, Baidu, and Tencent, along with numerous experts and scholars. This cross-industry collaboration aims to ensure that AI, particularly generative AI and large language models, develop securely and ethically, safeguarding user privacy and data security.
The “Safety Testing Standard for Generative Artificial Intelligence Applications” aims to provide a comprehensive safety assessment framework for AI system design and development, preventing potential misuse and deception. On the other hand, the “Safety Testing Method for Large Language Models” focuses on the training and application processes of language models, guaranteeing reliability and fairness when handling sensitive information and interactions.
The establishment of these international standards reflects the tech industry’s global emphasis on AI safety and the balance struck by companies between technological innovation and social responsibility. As AI technology becomes increasingly widespread, these standards will serve as vital references for developers, enterprises, and policymakers worldwide, fostering the healthy development of AI technology and the construction of a global digital ecosystem.
【来源】https://www.chinastarmarket.cn/detail/1649234
Views: 1