在瑞士日内瓦召开的第27届联合国科技大会上，一项重大科技合作成果正式揭晓。世界数字技术院（WDTA）宣布发布了《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》两项国际标准，旨在规范和保障人工智能领域的健康发展。这一里程碑式的事件标志着全球科技巨头对AI安全的集体关注和责任担当。
该标准的编制工作由OpenAI、蚂蚁集团、科大讯飞、谷歌等业界领先企业，以及微软、英伟达、百度、腾讯等数十家单位的专家学者共同参与完成。这些标准的出台，不仅体现了各企业在AI技术研发和应用中的深度合作，也彰显了他们在保障人工智能安全、隐私保护和伦理道德方面的共同承诺。
《生成式人工智能应用安全测试标准》聚焦于AI生成内容的可信度和可靠性，以防止误导性信息的传播。而《大语言模型安全测试方法》则旨在确保大规模语言模型在处理敏感信息和复杂对话时的准确性和安全性，防止技术滥用和潜在风险。
此次发布的国际标准将为全球AI行业的规范化发展提供有力指导，促进技术创新与社会安全的平衡，为人工智能的广泛应用奠定坚实的基础。同时，这也预示着全球科技界在面对AI带来的挑战时，正逐步形成统一的规范和标准，以实现技术进步与社会福祉的和谐共生。
英语如下：
News Title: “Global Tech Giants Unite to Launch International Standards for Large Model Safety at the UN Science and Technology Conference”
Keywords: Large Model Safety, International Standards, Science and Technology Conference
News Content: At the 27th United Nations Science and Technology Conference held in Geneva, Switzerland, a major milestone in tech collaboration was unveiled. The World Digital Technology Academy (WDTA) announced the release of the “Generative Artificial Intelligence Application Safety Testing Standard” and the “Large Language Model Safety Testing Method,” two international standards aimed at regulating and ensuring the healthy development of the AI sector.
This landmark event signifies the collective attention and responsibility of global tech giants towards AI safety. The standard development involved leading enterprises such as OpenAI, Ant Group, iFlytek, Google, and experts from Microsoft, NVIDIA, Baidu, Tencent, and several other entities. These standards not only demonstrate the depth of cooperation among companies in AI R&D and application but also underscore their shared commitment to AI safety, privacy protection, and ethical principles.
The “Generative Artificial Intelligence Application Safety Testing Standard” focuses on the credibility and reliability of AI-generated content to prevent the spread of misleading information. Meanwhile, the “Large Language Model Safety Testing Method” aims to ensure the accuracy and security of large language models when handling sensitive data and complex conversations, guarding against misuse and potential risks.
These newly released international standards will provide strong guidance for the standardized development of the global AI industry, fostering a balance between technological innovation and societal safety. They lay a solid foundation for the widespread application of AI. Furthermore, this signals a gradual convergence of norms and standards within the global tech community in response to the challenges posed by AI, paving the way for harmonious coexistence between technological progress and societal well-being.
【来源】https://www.chinastarmarket.cn/detail/1649234
Views: 2