《傅莹在创新经济论坛上关于人工智能的发言》

2020-04-14
导读
2019年11月21-22日,创新经济论坛在北京雁栖湖国际会议中心举行,其中“理解人工智能对人类的挑战”分组会由沃尔特·艾萨克森主持,亨利·艾尔弗雷德·基辛格博士、傅莹、埃克·施密特作为嘉宾发言。本文为傅莹的发言内容,整理自她的笔记。

【文/ 傅莹】

众所周知,人类距离实现超级人工智能还很遥远,然而,人工智能在一些具体领域和特定情况与约束之下,已经超越了人类,而且其范围在迅速扩大。人们对由此可能获得的好处寄予厚望,但同时,恐惧和担忧也随之而来。

亨利·基辛格博士谈到人工智能彻底改变人类意识的潜在可能性,担忧人工智能或将终结启蒙运动以来的理性时代。他是从历史、哲学和战略的角度看待这个问题的,他强有力的阐述令人深思。

美国在人工智能技术创新上处于领先地位,而中国随后,尤其在人工智能技术的大规模和活跃的应用方面成绩斐然,中美两国有更大的责任,去思考未来、思考应当怎么做。

但是,我们在谈论未来和如何面对技术进步之前,首先需要想明白,中美是要协调合作还是彼此对抗?当前两国之间日益恶化的紧张关系,必然会影响到我们如何应对未来的挑战。也就是说,未来我们是要共同努力,让技术与人类共生,让世界避免技术风险,并确保技术的进步能够促进文明的繁荣?还是要分道扬镳,各自挟持技术削弱甚至伤害对方?

我们的选择将影响到未来如何应对科技进步带来的挑战。冷战期间,美国和苏联经历了大小多次危机,有的甚至威胁毁灭人类,才最终达成某些自我约束和共存的安排。当今世界更加复杂,得失的影响更大,难道还需要更大的危机才能找到正确的道路吗?中美能否在现有世界体系内解决分歧、和平共处?抑或要如同美国一股力量所推动的那样,彼此脱钩,进而分裂世界?后者也是许多亚洲领导人乃至联合国秘书长等多方呼吁、警告的情形。

中国三十年多来一直处于高速工业化的进程中,采用的是两个世纪以来世界上陆续出现的所有基本规则。今天,中国第一次跻身新技术进步的第一梯队,除了尽己所能地向前迈进,中国人也逐渐意识到自身肩负的制定新规则的责任。

传统的立法方式是,要在社会形成共识的基础上,考虑如何制定规则。但是,现今的人工智能技术一旦投入应用,随即便会出现对恰当治理的需求。所以,对任何政府而言,人工智能技术应用的监管都是一项新挑战。对此,中国政府的政策是,一方面鼓励相关产业的发展,另一方面提供总体指导方针。

中国的国家新一代人工智能治理专业委员会于今年2月由科技部牵头组建,于6月发布新一代人工智能治理的八项原则,包括和谐友好、公平公正、包容共享、尊重隐私、安全可控、共担责任、开放协作、敏捷治理。为了推动这些原则落地,到2023年,政府将在全国布局建设20个新一代人工智能创新发展试验区,检验八项原则的实施和收集反馈。其他措施还包括提供开放平台,鼓励企业制定自己的标准,人工智能研发项目也需要遵守这八点原则。

今年8月,全国人大常委会审议通过民法典人格权编草案,其中针对收集、处理自然人个人信息做出明确规定,要求必须“征得该自然人或者其监护人同意”等,而且把个人生物识别信息也纳入保护范围。另一项重要进展是国家网信办发布的《儿童个人信息网络保护规定》,自2019年10月1日起生效。

人工智能技术的成长依靠的是科研人员分享思想、相互借鉴,是全球协作的产物,跨国企业构建的人工智能平台也在快速扩展。要想规范这个进程,各国需要制定互通的道德规范和行业规则。因此,中国在这方面的努力,也需要与世界其他国家和地区、包括美国的努力,相互联通和协调。为此,中国对与各国探讨扩大共识、缩小差异,秉持开放态度。

不论中国还是美国都不可能垄断世界的技术进步。如果两国采取互补的态度,人工智能技术的前景会更加光明;但如果不再合作,双方都将遭受损失,人工智能的总体发展也会付出代价。尤其是,如果任由传统的地缘政治、零和竞争思维主导两国关系,结果将是自毁性质的。

在增进了解和减少误解方面,中方可以做的努力包括,更加主动地与国际社会沟通。考虑到今日之中国在很大程度上处在世界舞台的聚光灯下,当我们发布一些重要文件时,其贴切的译文的发布同等重要,对误解的及时澄清也是必要的。例如,美方认为中国的野心是主导人工智能的未来,这方面的恐惧心理是美将科技视为与中国进行战略争夺平台的主要原因。而引发这种担忧的原因之一是,误读了2017年7月在北京发布的《新一代人工智能发展规划》。

规划提出了政府希争取实现的目标,第一步是到2020年中国的人工智能总体技术和应用与世界先进水平同步;第二步是到2025年基础理论实现重大突破,部分技术与应用达到世界领先水平;第三步是到2030年人工智能理论、技术与应用总体达到世界领先水平,成为世界主要人工智能创新中心。需要提示的是,规划中提出的目标是成为世界创新中心之一,而不是唯一和排他性的“中心”。对这个目标的正确理解非常重要,中方的这些期望都是合理和正常的。

美国试图把高科技作为战略争夺的平台,而中方对此并不以为然。实际情况是,在这一领域存在着建设性和战略性的相互依存,当然,无可否认,在科学和产业界,竞争在所难免。科睿唯安(Clarivate Analytics)的数据显示,从2013年到2017年,中美两国国际合作论文数量增长最快,合作论文达4000多篇。当前,美国企业在技术上(尤其半导体)领先,美国的大学也在世界上居于前列。而中国拥有最大的用户市场,为算法更快的迭代升级提供了条件。中美如能相得益彰,彼此都能从中受益,但如果美方执意推动脱钩,则会迫使中方寻求其他合作伙伴,或者自己设法解决,这也会削弱美国企业的地位和影响。

中国希冀的未来世界是一个相互依存的命运共同体,采取的政策是促进广泛国际对话,积极参与合作,鼓励制定共同规则,以实现安全、可靠、负责任的人工智能。正如习近平主席所指出的:“中国愿同国际社会一道,共创智能时代,共享智能成果。”

(作者系外交部前副部长、清华大学战略与安全研究中心主任。)

Fu Ying’s Remarks onUnderstanding the AI Challenge to Humanityduring the New Economy Forum

(A sub-Forum on Understanding the AI challenge to Humanity was held during the New Economic Forum in Beijing at 10:45am November 21, 2019, Yanqi Lake International Conference Center, Beijing. It was chaired by Walter Isaacson, participated by Dr. Henry Alfred Kissinger,Fu Ying, Eric Schmidt. The following is Fu Ying’s remarkes taken from her notes.)

Fu Ying:

We all know that the super-intelligent artificial intelligence (AI) is still far to reach for the mankind, however, AI has exceeded humans in some concreat domains and under specific constrains, and the areas are spreading rapidly. There is huge expectation of the benefit it may bring, but fear and concern are also rising.

Dr. Kissinger has talked about AI’s potential to fundamentally alter human consciousness and the danger of AI ending the age of reasoning since the Enlightenment. He is looking at it from a historical, philosophical and strategic perspective providing a powerful argument which has given us a lot to think about.

US is leading in AI technological innovation and China is following and is rapidly creating active and massive applications. These two countries should shoulder greater responsibilities in thinking ahead about the future and about what need to be done.

But before we can talk about the future, and about how we are going forward about technological advance, we first and foremost need to think through about whether we are going to do it together or if we are turning each other into enemies? The current worsening tension between China and the US cannot but have an impact on how we are going to deal with the challenges down the road.

That is to say, are we going to work together to make technology symbiotic to humans, help the world avoid technology risks, and ensure that the technology advances make our civilizations prosper? Or are we going to go separate ways and each trying to use the technology to undermine, even hurt the other side?

The choices we make will affect how we tackle the challenges posed by the advancement of science and technology. The US and the USSR during the cold war had to go through many crises,some threatened to bring the mankind to doom, before arriving at some kind of arrangement for self-constraint and co-existence.

The world today is more complicated and there is more at stake. Do we need bigger crisis to help us finding the right path? Can China and the US solve the differences and peacefully co-exist in the current world system or would they decouple like some people in Washington are trying to push for and therefore tearing the world apart, which is a prospect many Asian leaders and the UN Secretary General have been warning us against?

As far as China is concerned, the country has been on the fast pace of industrialization over the past three decades and has embraced all the basic rules developed in the past two centuries. Now, it is for the first time that China itself is among the first echelon in advanced technology. Apart from trying to do well, China is also growingly aware of rule-making responsibilities.

The traditional legislative approach is to wait for consensus in the society before a regulatory idea is developed. With current AI technology, the need for proper governance appears as soon as the application is made.

Therefore, it’s a new challenge for any government to manage this new regulatory challenge. The policy of the Chinese Government is on the one hand to encourage the advancement of industries, on the other hand, providing general guidelines.

China’s New Generation AI Governance Expert Committee which was set up by the Ministry of Science and Technology February 2019 has launched the 8 points principles in July. They include: harmony and human-friendliness, fairness and justice, inclusiveness and sharing, respect for privacy, security and controllability, shared responsibility, open collaboration and agile governance.

To put these principles into practice, the government will set up 20 pilot zones by 2023, for testing and collecting feedbacks on how the principles are being practiced. Other measures include offering open platforms to encourage enterprises to formulate their own standards. The R&D projects will also be measured by the 8 points principles.

In the Civil Code Part on Personality Rights which was adopted by the Standing Committee of the National People’s Congress August this year, there is particular emphasis on the protection of personal information (the collection and use of natural persons’ personal information can only be made with the consent of the natural person or the guardian). It also puts personal biometric data under protection. Another important development is the adoption by the National Network Information Office of the Provisions on the Cyber Protection of Children’s Personal Information which came into effective on October 1st, 2019.

AI technology research has always developed based on global collaboration, with researchers sharing ideas, and building on each other’s work. And AI platforms by multinational companies are also expanding globally. To regulate its progress, countries need to develop compatible ethical norms and industrial rules. So China’s governance effort needs to be connected with similar efforts in other parts of the world, the US in particular. The country also takes an open attitude in terms of discussing and working with other countries on creating and expanding common ground and reduce differences.

Neither China nor the US can monopolize the world’s technological progress. If the two countries complement each other, the prospects for AI technology is brighter; but if they stop working with each other, both will suffer and the general progress will pay a price. Should they allow the geopolitical and zero-sum competitive philosophy dominate their relations, it would be self-destructive.

One thing the Chinese side can do to improve the perceptions and reduce misunderstandings is to be proactive in communicating with the international community. Since China now is very much under the spot light of the world stage, it is important that when we launch major domestic documents, there should also be proper translation to go with them so that the world reads China’s intention accurately. Should there be misunderstandings, timely explanation is also necessary.

Take for example, the American fear of China’s ambition to dominate the AI’s future has prompted it to take the scientific and technological advance as the platfore of strategic rivalry. Aparantly its fear partly comes from mis-reading of the “New Generation Artificial Intelligence Development Plan” issued in Beijing July 2017.

The Plan laid out the targets the government wanted to promote, including: 1st step, for the Chinese AI technological development reaching the world advanced level by 2020; 2nd step, achieving major break through in basic theary and reaching the world leading level by 2025 in some parts of the technology and application; 3rd step, move to the forefront in theary, technology and application and become a major innovation center by 2030. It needs to be noted that what China wants to achieve is to become one of the world’s innovative centers, not “the” only and exclusive center. These are quite reasonable expectations.

The US is trying taking hi-tech as a platform of strategic rivalry which is not how China sees it. The reality in the field is a kind of constructive and strategic mutual dependency though no one can deny that competition in the fields of science and industry is not an abnormal phenomenon.

According to Clarivate Analytics, from 2013 to 2017, the number of AI related papers co-authored by Chinese and Americans grew the fastest, reaching 4,000 in 5 years. American companies lead in technologies (especially in semiconductor), and American universities are ahead of others in the world.  But China has the largest user market and therefore provides faster iterative upgrading for algorithm. So, the two countries can benefit tremendously in complementary partnership, unless the US forces a decoupling, and pushes China to find other partners or develop its own solutions which will also weaken Amercian companies’ position and influence.

So for China, the preferred future is a world of interdependent community with shared future and the policy is to engage in wide international conversations for encouraging collaboration and develop common rules for safe, reliable and responsible AI. As the Chinese President Xi Jinping said, “China is willing to work with the international community to create an age of intelligence and share the achievements of the intelligence.” (end)

Fu Ying is former Vice Minister of Foreign Affairs of China and the Chairperson of the Center for International Security and Strategy of Tsinghua University

上一篇:傅莹谈为什么美强硬势力要极力毒化中美合作气氛

下一篇:《把握变局、做好自己,迎接新的全球时代》