傅瑩在創新經濟論壇上關於人工智能的發言

【文/ 傅瑩】

眾所周知，人類距離實現超級人工智能還很遙遠，然而，人工智能在一些具體領域和特定情況與約束之下，已經超越了人類，而且其範圍在迅速擴大。人們對由此可能獲得的好處寄予厚望，但同時，恐懼和擔憂也隨之而來。

亨利·基辛格博士談到人工智能徹底改變人類意識的潛在可能性，擔憂人工智能或將終結啓蒙運動以來的理性時代。他是從歷史、哲學和戰略的角度看待這個問題的，他強有力的闡述令人深思。

美國在人工智能技術創新上處於領先地位，而中國隨後，尤其在人工智能技術的大規模和活躍的應用方面成績斐然，中美兩國有更大的責任，去思考未來、思考應當怎麼做。

但是，我們在談論未來和如何面對技術進步之前，首先需要想明白，中美是要協調合作還是彼此對抗？當前兩國之間日益惡化的緊張關係，必然會影響到我們如何應對未來的挑戰。也就是説，未來我們是要共同努力，讓技術與人類共生，讓世界避免技術風險，並確保技術的進步能夠促進文明的繁榮？還是要分道揚鑣，各自挾持技術削弱甚至傷害對方？

我們的選擇將影響到未來如何應對科技進步帶來的挑戰。冷戰期間，美國和蘇聯經歷了大小多次危機，有的甚至威脅毀滅人類，才最終達成某些自我約束和共存的安排。當今世界更加複雜，得失的影響更大，難道還需要更大的危機才能找到正確的道路嗎？中美能否在現有世界體系內解決分歧、和平共處？抑或要如同美國一股力量所推動的那樣，彼此脱鈎，進而分裂世界？後者也是許多亞洲領導人乃至聯合國秘書長等多方呼籲、警告的情形。

中國三十年多來一直處於高速工業化的進程中，採用的是兩個世紀以來世界上陸續出現的所有基本規則。今天，中國第一次躋身新技術進步的第一梯隊，除了盡己所能地向前邁進，中國人也逐漸意識到自身肩負的制定新規則的責任。

傳統的立法方式是，要在社會形成共識的基礎上，考慮如何制定規則。但是，現今的人工智能技術一旦投入應用，隨即便會出現對恰當治理的需求。所以，對任何政府而言，人工智能技術應用的監管都是一項新挑戰。對此，中國政府的政策是，一方面鼓勵相關產業的發展，另一方面提供總體指導方針。

中國的國家新一代人工智能治理專業委員會於今年2月由科技部牽頭組建，於6月發佈新一代人工智能治理的八項原則，包括和諧友好、公平公正、包容共享、尊重隱私、安全可控、共擔責任、開放協作、敏捷治理。為了推動這些原則落地，到2023年，政府將在全國佈局建設20個新一代人工智能創新發展試驗區，檢驗八項原則的實施和收集反饋。其他措施還包括提供開放平台，鼓勵企業制定自己的標準，人工智能研發項目也需要遵守這八點原則。

今年8月，全國人大常委會審議通過民法典人格權編草案，其中針對收集、處理自然人個人信息做出明確規定，要求必須“徵得該自然人或者其監護人同意”等，而且把個人生物識別信息也納入保護範圍。另一項重要進展是國家網信辦發佈的《兒童個人信息網絡保護規定》，自2019年10月1日起生效。

人工智能技術的成長依靠的是科研人員分享思想、相互借鑑，是全球協作的產物，跨國企業構建的人工智能平台也在快速擴展。要想規範這個進程，各國需要制定互通的道德規範和行業規則。因此，中國在這方面的努力，也需要與世界其他國家和地區、包括美國的努力，相互聯通和協調。為此，中國對與各國探討擴大共識、縮小差異，秉持開放態度。

不論中國還是美國都不可能壟斷世界的技術進步。如果兩國採取互補的態度，人工智能技術的前景會更加光明；但如果不再合作，雙方都將遭受損失，人工智能的總體發展也會付出代價。尤其是，如果任由傳統的地緣政治、零和競爭思維主導兩國關係，結果將是自毀性質的。

在增進了解和減少誤解方面，中方可以做的努力包括，更加主動地與國際社會溝通。考慮到今日之中國在很大程度上處在世界舞台的聚光燈下，當我們發佈一些重要文件時，其貼切的譯文的發佈同等重要，對誤解的及時澄清也是必要的。例如，美方認為中國的野心是主導人工智能的未來，這方面的恐懼心理是美將科技視為與中國進行戰略爭奪平台的主要原因。而引發這種擔憂的原因之一是，誤讀了2017年7月在北京發佈的《新一代人工智能發展規劃》。

規劃提出了政府希爭取實現的目標，第一步是到2020年中國的人工智能總體技術和應用與世界先進水平同步；第二步是到2025年基礎理論實現重大突破，部分技術與應用達到世界領先水平；第三步是到2030年人工智能理論、技術與應用總體達到世界領先水平，成為世界主要人工智能創新中心。需要提示的是，規劃中提出的目標是成為世界創新中心之一，而不是唯一和排他性的“中心”。對這個目標的正確理解非常重要，中方的這些期望都是合理和正常的。

美國試圖把高科技作為戰略爭奪的平台，而中方對此並不以為然。實際情況是，在這一領域存在着建設性和戰略性的相互依存，當然，無可否認，在科學和產業界，競爭在所難免。科睿唯安（Clarivate Analytics）的數據顯示，從2013年到2017年，中美兩國國際合作論文數量增長最快，合作論文達4000多篇。當前，美國企業在技術上（尤其半導體）領先，美國的大學也在世界上居於前列。而中國擁有最大的用户市場，為算法更快的迭代升級提供了條件。中美如能相得益彰，彼此都能從中受益，但如果美方執意推動脱鈎，則會迫使中方尋求其他合作伙伴，或者自己設法解決，這也會削弱美國企業的地位和影響。

中國希冀的未來世界是一個相互依存的命運共同體，採取的政策是促進廣泛國際對話，積極參與合作，鼓勵制定共同規則，以實現安全、可靠、負責任的人工智能。正如習近平主席所指出的：“中國願同國際社會一道，共創智能時代，共享智能成果。”

（作者系外交部前副部長、清華大學戰略與安全研究中心主任。）

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

本文系觀察者網獨家稿件，未經授權，不得轉載。