来源：AI与法
发布时间：2024-05-18
【编者按】2024年4月29日，美国商务部宣布根据美国总统在2023年12月颁布的《关于发展安全、可靠和值得信赖的人工智能的行政命令》，通过其下属的美国国家标准与技术研究院制定并发布四份指导性文件草案：促进生成式人工智能风险管理的《人工智能风险管理框架--生成人工智能档案》（NIST AI 600-1）、《生成人工智能和两用基础模型的安全软件开发实践》（美国国家标准与技术研究院特别出版物（SP）800-218A），规范合成内容风险管理的《降低合成内容带来的风险》（NIST AI 100-4），以及推动制定全球性人工智能标准的《人工智能标准全球参与计划》（NIST AI 100-5）。
同时，美国国家标准
与技术研究院针对生成式人工智能启动了一项名为“
NIST GenAI
”、旨在帮助人们识别和区分人工智
能生成内容和人类创作内容、评估和衡量生成式人工智能技术的能力和局限性的项目。
此外，美国商务部通过其下属的专利商标局在美国《联邦公报》发布指导意见以
明确
人工
智能辅
助发明的
可专利性
。
NIST发布的上述四份草案每份大概20多页，有兴趣阅读原文的小伙伴请移步文末参考文献[2]链接。
Department of Commerce Announces New Actions to Implement President Biden’s Executive Order on AI
美国商务部宣布新举措以落实拜登总统的人工智能行政命令
FOR IMMEDIATE RELEASE Monday, April 29, 2024
2024
年
4
月
29
日星期一即时发布
Announcements include draft guidance documents, a draft plan for international standards, and a new measurement program opening for public comment
新举措包括起草指导文件、起草一份国际标准计划和一个新措施项目供公开征求意见。
The U.S. Department of Commerce announced today, following the 180-day mark since President Biden’sExecutive Order (EO) on the Safe, Secure and Trustworthy Development of AI, several new announcements related to the EO. The Department’s National Institute of Standards and Technology (NIST) has released four draft publications intended to help improve the safety, security and trustworthiness of artificial intelligence (AI) systems. NIST has also launched a challenge series that will support development of methods to distinguish between content produced by humans and content produced by AI. In addition to NIST’s publications, Commerce’s U.S. Patent and Trademark Office (USPTO) is publishing a request for public comment (RFC) seeking feedback on how AI could affect evaluations of how the level of ordinary skills in the arts are made to determine if an invention is patentable under U.S. law, and earlier this year released guidance on the patentability of AI-assisted inventions.
美国商务部今日宣布，在拜登总统发布《关于发展安全、可靠和值得信赖的人工智能的行政命令》（EO，简称“《行政命令》”）满180天后，宣布几项与《行政命令》有关的新举措。该部下属的
国家标准与技术研究院（NIST）发布了四份文件草案，旨在帮助提高人工智能系统的安全性、可靠性和可信度。
美国国家标准与技术研究院还推动一系列挑战，以
支持开发区分人类创造内容和人工智能生产内容的方法
。除了美国国家标准与技术研究院发布的四份文件外，美国商务部下属的
专利商标局（USPTO）还发文公开征求意见（RFC），寻求对人工智能如何影响艺术领域普通技能水平评估的反馈，从而判断一项发明是否符合美国法律规定的专利申请
，该局还在今年早些时候发布了人工智能辅助发明专利申请指南。
“In the six months since President Biden enacted his historic Executive Order on AI, the Commerce Department has been working hard to research and develop the guidance needed to safely harness the potential of AI, while minimizing the risks associated with it,” said
U.S. Secretary of Commerce Gina Raimondo.
“The announcements we are making today show our commitment to transparency and feedback from all stakeholders and the tremendous progress we have made in a short amount of time. With these resources and the previous work on AI from the Department, we are continuing to support responsible innovation in AI and America’s technological leadership.”
美国商务部长吉娜·雷蒙多（
Gina Raimondo
）表示：“自拜登总统颁布历史性的人工智能行政命令以来的六个月时间里，商务部一直在努力研究和制定安全利用人工智能潜力同时将相关风险降至最低所需的指导意见。”“我们今天发布的公告证明，我们高度关注透明度和所有利益相关者的反馈，以及我们在短时间内取得的巨大进展。有了这些资源和国防部之前在人工智能方面的工作，我们将继续支持人工智能的负责任创新和美国的技术领导力。”
The NIST publications cover varied aspects of AI technology: The first two are guidance documents designed to help manage the risks of generative AI — the technology that enables chatbots and text-based image and video creation tools — and serve as companion resources to NIST’s AI Risk Management Framework (AI RMF) and Secure Software Development Framework (SSDF), respectively. A third NIST publication offers approaches for promoting transparency in digital content, which AI can alter; the fourth proposes a plan for developing global AI standards. These publications are initial drafts, which NIST is publishing now to solicit public feedback before submitting final versions later this year.
美国国家标准与技术研究院发布的这四份文件涵盖了人工智能技术的各个方面：前两份是指导性文件，旨在帮助管理支持聊天机器人和文生图和文生视频创建工具的生成式人工智能的风险，并分别作为美国国家标准与技术研究发布的《人工智能风险管理框架》（
AI RMF
）和《安全软件开发框架》（
SSDF
）的配套规则。美国国家标准与技术研究的第三份文件提出了提升人工智能可改变的数字内容的透明度的方法；第四份文件提出了制定全球人工智能标准的计划。美国国家标准与技术研究院目前发布这些文件初稿，旨在征求公众意见，今年晚些时候将提交最终版本。
“For all its potentially transformative benefits, generative AI also brings risks that are significantly different from those we see with traditional software,” said
Under Secretary of Commerce for Standards and Technology and NIST Director Laurie E. Locascio.
“These guidance documents will not only inform software creators about these unique risks, but also help them develop ways to mitigate the risks while supporting innovation.”
负责标准与技术业务的商务部副部长兼国家标准与技术研究院院长罗瑞·E·罗卡西欧（
Laurie E.Locascio
）表示：“尽管生成式人工智能具有潜在的变革性优势，但它也带来了与传统软件截然不同的风险。”“此次发布的指导文件不仅将揭示软件创作者所面临的这些独特风险，还将在支持创新的同时帮助他们开发出降低风险的方法。”
USPTO is publishing an RFC seeking feedback on how AI could affect evaluations they make as they determine whether an invention is patentable under U.S. law. For example, the use of AI poses questions as to what qualifies as prior art and the assessment of the level of skill of a person having ordinary skill in the art. USPTO expects the responses to the RFC will help them evaluate the need for further guidance on these matters, aid in the development of any such guidance, and help inform USPTO’s work in the courts and in providing technical advice to Congress.
美国专利商标局正在就人工智能如何影响他们判断一项发明是否符合美国法律所规定专利条件时所开展的评估征求意见。例如，人工智能的应用引发了什么是现有技术以及对于拥有本领域普通技能的人的技能水平的评估问题。美国专利商标局期待征求意见所获得的反馈能有助于他们评估就这些问题提供进一步指导的必要性、制定这类指导、引导美国专利商标局在法院的工作以及向国会提供技术建议。
“As AI assumes a larger role in innovation, we must encourage the responsible and safe use of AI to solve local and world problems and to develop the jobs and industries of the future, while ensuring AI does not derail the critical role IP plays in incentivizing human ingenuity and investment,” said
Under Secretary of Commerce for Intellectual Property and Director of the USPTO Kathi Vidal.
“This work builds on our inventorship guidance, which carefully set forth when the USPTO will issue a patent for AI-assisted innovations, and our continuing policy work at the intersection of AI and all forms of IP.”
负责知识产权业务的商务部副部长兼美国专利商标局局长凯西·维达尔（
Kathi Vidal
）表示：“随着人工智能在创新中发挥的作用越来越大，我们必须鼓励负责任、安全地使用人工智能来解决国内和国际问题，发展未来的就业和产业，同时确保人工智能不会破坏知识产权在激励人类创造力和投资方面发挥的关键作用。”“这项工作建立在我们的发明人指南的基础上，该指南详细规定了美国专利商标局在什么情况下会为人工智能辅助创新授予专利，以及我们在人工智能和所有形式的知识产权交叉领域的持续性政策工作。”
More information on the announcements being made today can be found below. All four of the NIST publications are initial public drafts, and NIST is soliciting comments from the public on each by June 2, 2024. Instructions for submitting comments can be found in the respective publications.
有关当日发布内容的更多信息，请参阅下文。美国国家标准与技术研究院发布的四份文件都是首次公开草案，它们将在2024年6月2日前就每一份文件征求公众意见。关于提交意见的说明，请参阅相应文件。
Mitigating the Risks of Generative AI
缓解生成式人工智能的风险
The
AI RMF Generative AI Profile
(NIST AI 600-1) can help organizations identify unique risks posed by generative AI and proposes actions for generative AI risk management that best aligns with their goals and priorities. Developed over the past year and drawing on input from the NIST generative AI public working group of more than 2,500 members, the guidance document is intended to be a companion resource for users of NIST’s AI RMF. The AI Profile guidance document centers on a list of 13 risks and more than 400 actions that developers can take to manage them.
《人工智能风险管理框架
--
生成人工智能档案》（
NIST AI 600-1
）可以帮助各组织识别生成式人工智能带来的特有风险，并提出最符合其目标和优先事项的生成式人工智能风险管理举措。
该指南是在过去一年中开发的，并借鉴了由2500多名成员组成的国家标准与技术研究院生成式人工智能公共工作组的意见，旨在成为国家标准与技术研究院《人工智能风险管理框架》用户的配套规则。
《生成式人工智能档案》指南围绕生成式人工智能的
13
种特有风险和开发者可采取的
400
多项行动为中心
。
The 13 risks include issues such as easier access to information related to chemical, biological, radiological or nuclear weapons; a lowered barrier to entry for hacking, malware, phishing, and other cybersecurity attacks; and the production of hate speech and toxic, denigrating or stereotyping content. Following the detailed descriptions of these 13 risks is a matrix of the 400 actions that developers can take to mitigate the risks.
这13种风险包括如下问题：更容易获得与化学、生物、放射性或核武器有关的信息；降低了黑客、恶意软件、网络钓鱼和其他网络安全攻击的攻击门槛；制造仇恨言论和有毒、诋毁或偏见内容。在详细描述这13种风险后，这份指南为开发人员提供了可以采取的400种降低风险的措施矩阵。
Reducing Threats to the Data Used to Train AI Systems
减少对人工智能系统训练所使用数据的威胁
The second publication,
Secure Software Development Practices for Generative AI and Dual-Use Foundation Models
(NIST Special Publication (SP) 800-218A), is designed to be used alongside the SSDF (SP 800-218). While the SSDF is broadly concerned with securing the software’s lines of code, this companion resource expands the SSDF to help address concerns around malicious training data adversely affecting generative AI systems.
第二份文件《生成人工智能和两用基础模型的安全软件开发实践》（美国国家标准与技术研究院特别出版物（
SP
）
800-218A
）旨在与《安全软件开发框架（
SSDF
）
1.1
版：降低软件漏洞风险的建议》（
Secure Software Development Framework (SSDF) Version 1.1: Recommendations for Mitigating the Risk of Software Vulnerabilities
）（
SP 800-218
）共同发挥作用
。虽然《安全软件开发框架（SSDF）广泛关注软件代码行的安全，但此次配套规则扩展了《安全软件开发框架（SSDF），
以帮助解决对生成式人工智能系统产生不利影响的恶意训练数据问题。
In addition to covering aspects of the training and use of AI systems, the new companion resource offers guidance on dealing with the training data and data collection process, including a matrix that identifies potential risk factors and strategies to address them. Among other recommendations, the document suggests analyzing training data for signs of poisoning, bias, homogeneity and tampering.
除了涵盖人工智能系统的培训和使用方面，新的配套资源还提供了处理培训数据和数据收集过程的指导，包括一个确定潜在风险因素和应对策略的矩阵。在其他建议中，该文件建议分析训练数据中的中毒、偏见、同质性和篡改迹象。
Reducing Synthetic Content Risks
降低合成内容的风险
Accompanying generative AI’s development has been the rise of “synthetic” content, which has been created or altered by AI. Offering technical approaches for promoting transparency in digital content is the goal of NIST’s new draft publication,
Reducing Risks Posed by Synthetic Content
(NIST AI 100-4). This publication informs, and is complementary to, a separate report on understanding the provenance and detection of synthetic content that AI EO Section 4.5(a) tasks NIST with providing to the White House.
伴随生成式人工智能发展的是人工智能创造或修改的“合成”内容的增加。
提高数字内容透明度的技术方法是美国国家标准与技术研究院最新发布的第三份文件草案《降低合成内容带来的风险》（NIST AI 100-4）的目标
。
该文件形成并补充了人工智能
《行政命令》第4.5条第（a）款规定美国国家标准与技术研究院应当向白宫提供的合成内容来源和检测的单独报告
。
NIST AI 100-4 lays out methods for detecting, authenticating and labeling synthetic content, including digital watermarking and metadata recording, where information indicating the origin or history of content such as an image or sound recording is embedded in the content to assist in verifying its authenticity. The report does not focus only on the dangers of synthetic content; it is intended to reduce risks from synthetic content by understanding and applying technical approaches for improving the content’s transparency, based on use case and context.
《降低合成内容带来的风险》（NIST AI 100-4）草案列出了
检测、验证和标记合成内容的方法
，包括数字水印和元数据记录，其中指示内容的来源或历史的信息（如图像或录音）嵌入内容中以帮助验证其真实性。该报告不仅
关注合成内容的危险性；也旨在通过理解和利用以用例和上下文为基础的技术方法来提高内容的透明度，从而降低合成内容的风险。
Global Engagement on AI Standards
人工智能标准的全球参与
AI systems are transforming American society and around the world.
A Plan for Global Engagement on AI Standards
(NIST AI 100-5) is designed to drive the worldwide development and implementation of AI-related consensus standards, cooperation and coordination, and information sharing.
人工智能系统正在改变美国社会和世界各地。美国国家标准与技术研究院最新发布的
第四份文件草案《人工智能标准全球参与计划》（NIST AI 100-5）旨在推动全球人工智能相关共识标准的制定和实施、合作与协调以及信息共享。
