我们,机器人? We, the Robots — in Chinese!
We, the Robots? Regulating Artificial Intelligence and the Limits of the Law has been translated into Chinese, with a new preface!
(点击此处立即购买！)
More than three decades after I lived in Beijing for a year, studying at 北京第二外国语学院 (now the Beijing International Studies University), it’s a thrill to have my first book published in Chinese. Alas my own language skills were not up to the task, so I’m deeply grateful to You Chuanman (游传满) and Fei Xiuyan (费秀艳) for their efforts. Translating legal texts is sufficiently difficult that it has its own term: jurilinguistics. Thank you also to the team at Peking University Press for such a distinctive look and feel, and to esteemed colleagues and friends Justice Xue Hanqin, Prof Guo Li, Prof Zeng Yi, and Prof Wang Jiangyu for their kind words in support of the new edition.
The reason for all this work was not merely so that the book might reach a wider audience. China is a vital and understudied actor in the AI governance space. In my own research, I have tried to reflect that importance. The pandemic made travel there difficult in recent years, though I was able to connect with many scholars and students remotely. More recently, I have returned to Beijing and will soon travel once more to Shanghai and beyond. I have learned a lot from these engagements and hope that having my written work more accessible will facilitate further discussions.
Though the translation is faithful to the original text, a new preface updates some of the content.
The book is available through Chinese bookstores below, or you can get the English version at the outlets here.
Advance Praise (专家推荐)
人工智能的技术革命将极大地解放劳动生产力，推动经济社会的快速发展。与以往技术革命不同的是，人工智能对人类社会现存形态的冲击，将是持续不断而深远难测的。我们能否对人工智能的发展进行有效的监管和控制是一个前所未有的挑战。相信陈西文教授的独到见解将给读者带来深刻的思考。
—- 薛捍勤 (Xue Hanqin, Justice of the International Court of Justice)
陈西文教授的新书将规制主体、法律责任与科技伦理等人工智能治理中的核心议题巧妙融合，辅之以证券交易、行政决策与自动驾驶等多个领域的生动案例，为主权国家积极参与、国际法及国际机制居中协调的全球人工治理方案勾画出一份蓝图。
—- 郭雳，北京大学法学院院长、教授 (Guo Li, Peking University Law School）
陈西文教授的新书为监管人工智能提供了详细的框架与分析。令我最感兴趣的是作为法学专家的陈教授清晰地指出在监管人工智能过程中法律的位置与局限。全书关于人工智能技术的讨论与监管框架的讨论在非常恰当的层次进行深度融合。他提出的终极问题，亦是本书的书名及我认为本书的精神内核，即是提醒我们人工智能在社会中应当置于恰当的位置，且我们绝对不需要使人工智能无处不在。
—- 曾毅，中国科学院自动化研究所研究员，人工智能伦理与治理研究中心主任，联合国教科文组织人工智能伦理特设专家组专家 (Zeng Yi, Chinese Academy of Sciences)
陈西文教授是世界范围内最早研究人工智能与法律的学者，可以说是这个交叉学科的奠基人之一。陈教授以广阔的视野和深邃的分析能力，前瞻性地探讨法律如何监管认知能力正在飞速发展的人工智能，他的这本书绝对是这个领域的必读和经典之作。
—- 香港城市大学法律学院教授 王江雨 (Wang Jiangyu, City University of Hong Kong)
Buy the Book in Chinese (购书链接)
Preface to the Chinese Edition (中译本序)
目前,人工智能的迭代发展可 谓日新月异。世界各国有识之士在探讨如何充分发挥人工智能价 值贡献的同时,也在努力探寻如何减少或降低人工智能对社会、 经济、政治、法律各方面带来的风险冲击。讨论的关注焦点多集 中在美国科技巨头的产业创新,比如谷歌的Gemini、微软和 OpenAI的ChatGPT,或者欧盟监管机构的制度创新,比如最近 通过的《人工智能法》(AI Act)。与此同时,中国企业在人工智 能模式创新与应用生态发展上也扮演着日益重要的作用,中国政 府在如何创造性地引导与规范人工智能方面也变得更加积极 有为。
特别是生成式人工智能的出现,使得这些讨论变得更具紧迫 性与现实性。过去,许多人认为人工智能的短期影响主要体现在 算法推荐系统和人脸图像识别上。而其中的突出风险体现在偏见 与歧视、无益建议以及身份盗用上。然而,2016年美国总统大选 中出现的各种问题,说明人工智能很可能在更大范围内被用来操 纵普罗大众。现如今,随着新型人工智能,诸如ChatGPT、 Gemini、Ernie Bot (文心一言)、Tongyi Qianwen (通义千问) 等大语言模型(Large Language Model, LLM)的产生与应用, 还出现了更多重要的治理规范命题。
其中,侵犯知识产权可能是它们的“原罪”,因为大语言模 型多是在海量数据上进行训练的,而这些海量数据就包括盗版侵 权作品,以及被分享以供阅读但未经授权复制的信息数据。这是 输入端的问题。除此之外,输出端的所有权问题也存在极大争 议。大多数国家目前还不承认计算机对其生成内容在知识产权法 律体系中的所有权,尽管中国最近的判例已经部分支持了此类所 有权。
这些法律规范问题可能在根本上影响到知识经济(knowledge economy)对于社会发展转型变革的影响。“知识工作者” (knowledge worker)是现代管理学之父彼得· 德鲁克(Peter Drucker)在1959年为非传统问题解决者创设的专有名词,他们 是“以思考为生”的职业人士,他们通过分析和创作的专业能力 赚取收入———此类分析与创作是ChatGPT 等大语言模型可以在 几乎零时间成本与零资金成本的情况下完成的。如今,此类职业 人士迫切地想知道大语言模型是否会在不久的将来取代他们的 工作。
鉴于以上诸种原因,我非常荣幸地向中文读者呈献我早前于 剑桥大学出版社出版的《我们,机器人? ———人工智能监管及其 法律局限》一书的中译本。此中译本的出版可谓正逢其时。我本 人长期在新加坡生活和工作,时常觉得本地关于人工智能治理的 讨论偏于狭隘和肤浅。所谓偏于狭隘,我的意思是讨论由少数声 音主导———特别是美国和西欧的声音,虽然偶尔也涉及中国。所 谓偏于肤浅,我的意思是讨论往往局限于治理原则等形而上学的 问题。
希望此中译本的面世,能为如何促进人工治理与监管更广 泛、更深入的讨论贡献我的绵薄之力。借由此中译本,我也衷心 期待能够与中文理论界和实务界同仁在人工智能领域进行更直接 的交流和探讨。正如我在前言部分的结尾所写的,拙著很可能无 法提供所有答案。但我希望它能帮助我们共同努力,以寻找和确 定问题所在。
我要借此机会,再次感谢在英文致谢中列出的同事与家人。 当然,此中译本能在这么短的时间内与中文读者见面,还需要感 谢游传满和费秀艳两位学者夜以继日的付出。特别是游传满博 士,早在英文版本的写作过程中,我们就对人工智能的司法应用 和数据保护等议题开展过探讨。
“借你们文采飞渡,此书得以横跨语言之桥!”
最后,我还要感谢北京大学出版社的孙维玲女士及其同事 们,以及剑桥大学出版社的伍志宏（Joe Ng）及其众多 同事们。
陈西文 (Simon Chesterman), 于新加坡
Around the world, countries are grappling with the question of how to reap the benefits of artificial intelligence while minimizing or mitigating the risks. Much of the attention has focused on the work of US technology giants like Google and Microsoft, or European regulators and the new AI Act. At the same time, Chinese companies are playing an increasingly important role in developing the computer systems that make AI possible, and its government has become more active in thinking creatively about how AI can and should be governed.
The emergence of generative AI has made these questions more urgent. In the past, many thought of the short-term impact of AI as being in recommendation systems and image recognition. Risks included bias, unhelpful suggestions, and perhaps a chance of identity theft. Then questions about the 2016 US Presidential election suggested that AI might be used to manipulate people on a far larger scale. Today, many employees wonder if large language models (LLMs) like ChatGPT, Gemini, Ernie Bot ( 文心一言), or Tongyi Qianwen (通义千问) might replace some or all of what they do at work.
These models raise important governance questions throughout their lifecycle. Copyright violation might be the “original sin” of AI, as most of the LLMs are trained on vast quantities of data that includes pirated works and material that was shared to be read but not copied. In addition to questions about their inputs, ownership of their output is disputed. Most countries do not recognize ownership of computer-generated content for the purposes of intellectual property law, though China has suggested that this may be possible.
These legal questions are at the heart of far larger social and economic changes that could affect the knowledge economy. “Knowledge workers” was the term introduced in 1959 by management consultant Peter Drucker for non-routine problem solvers. People who “think for a living” earn through their ability to analyse and write — something that ChatGPT and so on can replicate in almost no time and at almost no cost.
For all these reasons, it is timely — and I am deeply honoured — to release a new Chinese edition of my book on the regulation of AI.
Based in Singapore, it often strikes me that discussions about AI governance are often too narrow and too shallow. By too narrow, I mean that a few voices dominate the discussion – notably those in the United States and Europe, with China sometimes included. By too shallow, I mean that discussion is often limited to generalities and principles.
I hope that this book helps encourage a wider and deeper discussion about governance and regulation of AI. In particular, I look forward to engaging with and learning from Chinese scholars and practitioners in this field, and hope that my own work may make a modest contribution to ongoing debates here and beyond.
In addition to those thanked in the original acknowledgements of my book, I wish to recognize the tremendous contributions of You Chuanman 游传满 and Fei Xiuyan 费秀艳 for their dedicated work in translating the work into Chinese. Through your literary talents, the book can cross the bridge of language. (借您们文采飞渡，此书得以横跨语言之桥.)
Thank you also to Sun Weiling and her colleagues at Peking University Press, as well as Joe Ong and his colleagues at Cambridge University Press.
As I write at the end of the introduction, a work such as this cannot hope to provide all the answers. But I hope it can help us all work together and focus on the right questions.
Simon Chesterman (陈西文), in Singapore
Further Reading (附加信息)
Read Professor Guo Li’s full comment on the book here.
For additional material in English, including reviews of the book, see this page.
