**全球巨头联手制定大模型安全国际标准，推动人工智能健康发展**
近日，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）公布了具有里程碑意义的国际技术标准。经过OpenAI、蚂蚁集团、谷歌、科大讯飞等全球科技巨头共同协作，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大国际标准正式出炉。这一标准的制定对于人工智能的健康发展具有深远影响。
随着生成式人工智能应用的日益普及，确保这些技术安全可靠变得至关重要。联合编制的专家团队深入探讨了如何评估人工智能模型的潜在风险，包括隐私泄露、偏见和歧视等问题。此次发布的国际标准旨在确保大模型在各种应用场景中的安全性，为用户带来更加可靠和透明的AI体验。这一重要成果得益于全球数十家科技企业的专家学者共同智慧与努力。标准的发布将为行业发展提供方向性指导，确保人工智能技术能够更好地造福人类。这是全球技术社群合作共进、促进人工智能技术可持续发展的标志性事件。百度、腾讯等知名科技巨头亦参与其中，共同书写人工智能安全的新篇章。
英语如下：
News Title: “Global Tech Giants Collaborate to Set International Standard for Large Model Security, Published at the UN Tech Conference”
Keywords: Large Model Security International Standard, Joint Development, Publication
News Content: **Global Tech Giants Collaborate to Set Standard for Large Model Security, Driving Healthy Development of AI**
At the 27th UN Conference on Technology held in Geneva, Switzerland, recently, the World Digital Technology Academy (WDTA) unveiled a milestone international technical standard. Through the joint efforts of OpenAI, Ant Group, Google, iFLYTEK, and other global tech giants, two international standards for generative AI application security testing and large language model security testing methods were officially launched. The formulation of this standard has a far-reaching impact on the healthy development of artificial intelligence.
With the increasing popularity of generative AI applications, ensuring their technical safety and reliability has become crucial. The jointly-formed expert team delved into how to assess the potential risks of AI models, including privacy breaches, biases, and discrimination. The newly published international standards aim to ensure the safety of large models in various application scenarios, delivering a more reliable and transparent AI experience to users. This significant achievement is the result of the joint wisdom and efforts of experts and scholars from dozens of technology companies worldwide. The publication of the standard will provide directional guidance for industry development and ensure that AI technology can better benefit humanity. This is a landmark event for global tech communities to collaborate and promote the sustainable development of AI technology. Baidu, Tencent, and other renowned tech giants were also involved, writing a new chapter in AI security together.
【来源】https://ai-bot.cn/go/?url=aHR0cHM6Ly93d3cuY2hpbmFzdGFybWFya2V0LmNuL2RldGFpbC8xNjQ5MjM0
Views: 1