【观点】国务院发展研究中心副主任、党组成员陈昌盛：数字经济应优先发展哪些要素？

近日，以“数字技术与实体经济融合赋能数字中国”为主题的第五届数字发展论坛在京举办。国务院发展研究中心副主任、党组成员陈昌盛以“关于当前促进数字经济发展的几点认识”为题作特邀演讲。

陈昌盛表示，数实融合与数字经济的发展存在三方面力量，分别是技术突破、商业创新和监管制度的力量。他认为，如何更好地发展数字经济，从政策层面看其实是一个权衡与取舍的问题，基于此，他提出了“六个优先”：一是数据的可及性优先于数据的确权；二是数据流通优先于数据交易；三是在数据监管中小步快跑要优先于看准了再干；四是公共数据的开放要远远优先于开发；五是个人数据的匿名化利用优先于确权分配；六是重要数据的正负面清单结合要优先于准确的安全评估。此外，他指出，平台企业的商业边界及其公共职能的界定同样值得重视。

我主要从宏观的角度谈一谈对数字经济发展的几点感受。大家都说人工智能发展有算力、算法、数据三股支撑性力量，我认为数实融合与数字经济的发展同样也存在三方面力量，分别是技术突破、商业创新和监管制度的力量。

这三方面力量中，过去两年更多聚焦在制度层面，大家普遍认为监管对数字经济发展产生了重大影响，而今年形势发生了一些积极变化。2022年11月，ChatGPT正式对外发布，生成式人工智能这种革命性的技术力量，对全球竞争格局产生了一定冲击，人工智能在过去被认为无法从事创造性工作，在科学研究中处于辅助、从属性地位，但当前，人工智能本身成为了创新者和科学家。这些变化使得今年大家对数字经济的讨论更侧重于技术和商业模式创新。当然，制度上也做了一些积极的回应，比如组建国家数据局，颁布“数据二十条”等，大家很关心的公共数据开放、常态化监管机制等问题也都有所进展，这都是制度层面传递出来的一些积极信号。所以无论是从技术突破、商业创新，还是从国家监管层面来看，数字经济发展都在回归正轨，这是非常令人欣喜的现象。 CXO UNION-CXO联盟（cxounion.cn）

如何更好地发展数字经济，从政策层面看其实是一个权衡与取舍的问题，是一个发展与安全的权衡，是一个新动能与旧势力的对抗和替代过程。权衡与取舍就涉及到优先序，基于此，我就当前的一些热点问题，在去年提出的数字宏观“八大趋势”的基础之上，再谈“六个优先”。

一是数据的可及性优先于数据的确权。社会各界对此已经达成了一定共识，2022年12月颁布的“数据二十条”淡化所有权，提出建立数据资源持有权、数据加工使用权和数据产品经营权“三权分置”的数据产权制度框架。数据交易的一大特点是，确权不是交易的前提。科斯定理告诉我们，如果没有交易成本，确权本身并不重要。很多法学家认为，交易的发生与责任的认定建立在数据确权的基础之上，但实践表明，即便在当前世界各国都没有对数据进行确权的情况下，仍然存在责任追究机制，而只要形成了较好的责任认定和惩奖机制，数据确权就不一定是必须环节。当前，不少地方政府发文称地方公共数据属于当地政府，这是需要警惕的，其中的权属问题还有待商榷，数据的可及性比确权更重要，要先把数据用起来。 CXO UNION-CXO联盟（cxounion.cn）

二是数据流通优先于数据交易。每当谈及数据，数据交易往往是各方关注的焦点，似乎数据只有发生过交易才能使用。数据交易当然是存在的，但是现实生活大量的数据使用不依赖于数据交易，例如数据在垂直领域的流通、在供应链上下游的流转等，数据交易只是数据流通方式的一种，只占数据流通总量很小的一部分。此外，在数据交易中大家关注更多的是场内交易，而没有关注到大量的场外交易，之所以说“数据交易所悖论”，数据交易所遍地开花，但数据交易所中的“交易”发展的并不是很好。市场主体之间自由的数据流通行为是大量存在的，在各地掀起的数据市场建设热潮下，应警惕形成“只有经过交易买卖的数据才是可用数据”的行为导向。 CXO UNION-CXO联盟（cxounion.cn）

三是在数据监管中小步快跑要优先于看准了再干。实践发展领先于理论，也领先于政策监管。创新行为本身是容易辨别的，但创新行为的影响是难以预料的，这容易造成政策激励机制的扭曲。对于不清楚的事项为什么要批呢？审批行为产生的相关责任如何划分？“小步快跑”说的是，对于某项事情，本来有九点看得懂，一点看不懂，就应当放九点而保留一点，而非整个十点都拖着不放。“小步快跑”的监管模式对于推动人工智能的发展尤为重要，人工智能的未来确实充满了不确定性，在比较有把握的方面应先让它“走走看”。在当前的政策激励机制下，决策主体以“免责”为目标，这样的激励机制存在一定扭曲。对于新兴事物，看准之后再做决策是不可能的，没有人能完全看得准，因此监管的“小步快跑”就变得尤为重要。

四是公共数据的开放要远远优先于开发。由于“土地财政”已不可持续，在部分地方上出现了从“土地财政”转向“数据财政”的苗头。但是我一直认为，在本质上，民众对公共数据已经支付了费用，无论是从国际惯例还是法理道义上，公共数据应优先开放。开放是最好的开发，公共数据开放后，来自社会各界的开发力量将大量涌现。政府仅依靠自身力量开发公共数据存在技术、经验等方面的短板，而政府授权企业开发公共数据，如果只选择一家国有企业进行授权，这很容易形成新一轮垄断。为什么只授权给一家？为什么不授权给民营企业？政府不少数据是机密的，但只要遵循一定的规制，国企民企的选择不应存在差别。在地方上，很多政府不愿、不敢或者不想把数据开放出来，要警惕过度重视“开发”可能会造成新一轮扭曲。 CXO UNION-CXO联盟（cxounion.cn）

五是个人数据的匿名化利用优先于确权分配。“数据二十条”颁布后，我们分析了互联网上几十万篇文章，有意思的是，个人对数据确权和利润分配的热情突然高了起来。但事实上“数据二十条”里讲的，公共数据、企业数据、个人数据三者不在同一维度，“数据二十条”也并非从所有权的角度展开论述的。从个人的角度出发，最重要的是保护个人隐私安全。这其中有很多复杂的问题，包括个人数据匿名化使用规则的制定，另有一些数据无法匿名化，例如个人的指纹、声纹、虹膜等，这类数据的使用应当遵照什么样的规则？欧盟已经颁布了一些可参考的规则，例如250人以下的企业适用于一些免责条款等，国内在这方面规则的空缺应当尽快补齐。

六是重要数据的正负面清单结合要优先于准确的安全评估。重要数据的安全评估涉及利益相关方多、评估难度大，什么是重要数据、用什么方法评估安全数据等问题难以明确，可采用正面清单和负面清单相结合的模式。2023年9月，国家网信办发布《规范和促进数据跨境流动规定（征求意见稿）》，正面列明了豁免清单，是数据流动规则制定的极大进步，但是其中仍存在很多问题，我个人的建议是“正面清单”或“负面清单”应看准一条放一条，不清晰的领域应在市场实践中形成答案，以避免不必要的企业合规成本。 CXO UNION-CXO联盟（cxounion.cn）

以上是透过新一轮技术突破和商业创新反观政策制度所产生的一些思考。此外，平台企业的商业边界及其公共职能的界定同样值得重视。当互联网平台发展到一定规模，就自然具备了一定的公共属性，在这个问题上平台企业非常困扰。互联网平台的公共属性该如何界定？依托于平台产生的什么数据属于公共数据？哪些数据应当拿出来服务于公众？这是当前很多平台所担心的，也是需要尽快明确的。

近期斯坦福大学做了一个关于人工智能的全球调查，其中包括各国居民对人工智能的态度。调查结果表明，中国人对人工智能的态度是全球最积极的，78%的中国人认为人工智能利大于弊，这一比例在美国只有35%，中国人乐于且勇于尝试新鲜事物，这种好奇心文化难能可贵。当前，数字经济领域还有的一些投资资金，有没有躺平的投资热情，我们应该格外珍惜，创造更好的制度，使监管、技术与商业之间的互动更畅通，保证在这新一轮数字技术或者数字经济的发展浪潮中，我们不被落下！ CXO UNION-CXO联盟（cxounion.cn）

翻译;

Chen Changsheng, Deputy Director of the Development Research Center of The State Council and member of the Party Group: What elements should be prioritized for the development of the digital economy?

A data property right system framework of data resource holding right, data processing right and data product management right is proposed. One of the peculiarity of data trading is that certainty is not a prerequisite for trading. Coase’s theorem tells us that if there are no transaction costs, certainty itself does not matter. Many jurists believe that the occurrence of transactions and the identification of responsibility are based on the confirmation of data rights, but practice shows that even in the current situation where countries in the world do not confirm data rights, there is still a responsibility investigation mechanism, and as long as a better responsibility identification and punishment and reward mechanism is formed, data confirmation of rights is not necessarily a necessary link. At present, many local governments have issued a document saying that local public data belongs to the local government, which needs to be vigilant, and the ownership issue is still to be discussed, the accessibility of data is more important than the confirmation of rights, and the data must be used first.

Second, data circulation takes precedence over data trading. When it comes to data, data trading is often the focus of attention, as if data can only be used if it has been traded. Data transaction certainly exists, but a large amount of data use in real life does not depend on data transaction, such as the circulation of data in the vertical field, the flow of data in the upstream and downstream of the supply chain, etc. Data transaction is only one kind of data circulation mode, accounting for only a small part of the total data circulation. In addition, in data trading, people pay more attention to the exchange trading, and do not pay attention to a large number of over-the-counter transactions, the reason why the “data exchange paradox”, data exchanges everywhere, but the development of “trading” in data exchanges is not very good. There is a large number of free data flow behaviors among market entities, and under the upsurge of data market construction set off around the country, we should be wary of forming the behavior orientation of “only the data that is traded is available data”.

The third is that in data supervision, walking fast should be prioritized over doing it again. Practice is ahead of theory and ahead of policy regulation. The innovation behavior itself is easy to identify, but the impact of innovation behavior is difficult to predict, which easily leads to the distortion of policy incentive mechanism. Why approve things that are not clear? How to divide the related responsibilities arising from the approval act? “Small steps” means that if you can understand something at nine points, and if you can’t understand it at all, you should put it at nine points and keep it at one point, instead of dragging it through the whole ten points. The regulatory model of “small step and fast run” is particularly important to promote the development of artificial intelligence, the future of artificial intelligence is indeed full of uncertainty, and it should be allowed to “walk and see” in the more certain aspects. Under the current policy incentive mechanism, the decision-making subject takes “exemption” as the goal, and such incentive mechanism has certain distortion. For emerging things, it is impossible to make decisions after seeing it correctly, no one can see it completely, so the “small step fast run” of regulation has become particularly important. CXO UNION-CXO联盟（cxounion.cn）

Fourth, the opening of public data should take precedence over development. Because “land finance” is no longer sustainable, in some places there is a shift from “land finance” to “data finance”. But I have always believed that, in essence, people have paid for public data, and that public data should be open first, both in terms of international practice and jurisprudence. Openness is the best development, after the opening of public data, the development force from all walks of life will emerge in large numbers. The government only relies on its own strength to develop public data there are shortcomings in technology, experience and other aspects, and the government authorizes enterprises to develop public data, if only a state-owned enterprise is selected for authorization, it is easy to form a new round of monopoly. Why only one? Why not license it to private enterprises? Much government data is confidential, but as long as certain regulations are followed, there should be no difference in the choice of state-owned and private enterprises. At the local level, many governments are unwilling, afraid or unwilling to open up their data, and we must be vigilant that excessive emphasis on “development” may cause a new round of distortions. CXO UNION-CXO联盟（cxounion.cn）

Fifth, the anonymization of personal data takes precedence over the allocation of rights. After the “Data Article 20” was promulgated, we analyzed hundreds of thousands of articles on the Internet, and it was interesting that individuals’ enthusiasm for data confirmation and profit distribution suddenly increased. However, in fact, the public data, enterprise data and personal data mentioned in the “Data 20” are not in the same dimension, and the “Data 20” is not discussed from the perspective of ownership. From an individual’s point of view, the most important thing is to protect personal privacy and security. There are many complex issues, including the formulation of rules for the use of anonymization of personal data, and some data that cannot be anonymized, such as personal fingerprints, voice prints, iris, etc., what rules should be followed for the use of such data? The EU has issued some rules for reference, such as some exemption clauses for enterprises with less than 250 employees, and the vacancy of domestic rules in this area should be filled as soon as possible.

Sixth, the combination of positive and negative lists of important data should take precedence over accurate safety assessments. The security assessment of important data involves many stakeholders and is difficult to evaluate. It is difficult to clarify what is important data and how to evaluate security data. Therefore, a model combining positive and negative lists can be adopted. In September 2023, the National Cyberspace Administration issued the Regulations on Regulating and Promoting Cross-border Data Flow (Draft for Comment), which listed the exemption list in a positive way, which is a great progress in the formulation of data flow rules, but there are still many problems. My personal suggestion is that the “positive list” or “negative list” should be placed one by one. Areas that are unclear should form answers in market practice to avoid unnecessary corporate compliance costs.

These are some of the reflections generated by reviewing the policy system through a new round of technological breakthroughs and business innovation. In addition, the commercial boundary of platform enterprises and the definition of their public functions are also worthy of attention. When the Internet platform develops to a certain scale, it naturally has certain public attributes, and platform enterprises are very troubled on this issue. How to define the public attributes of Internet platform? What data generated on the platform is public data? What data should be made available to the public? This is what many platforms are worried about, and it needs to be clarified as soon as possible.

Recently, Stanford University conducted a global survey on artificial intelligence, including the attitudes of people in various countries towards artificial intelligence. The survey results show that Chinese people’s attitude toward artificial intelligence is the most positive in the world, 78% of Chinese people believe that artificial intelligence has more advantages than disadvantages, this proportion is only 35% in the United States, Chinese people are willing and brave to try new things, this kind of curiosity culture is valuable. At present, there are still some investment funds in the field of digital economy, there is no flat investment enthusiasm, we should cherish it, create a better system, make the interaction between regulation, technology and business more smooth, and ensure that in this new round of digital technology or digital economy development wave, we are not left behind!

由CXO UNION-CXO联盟（cxounion.cn）转载而成，来源于伏羲智库；编辑/翻译：CXO UNIONCXO联盟小U。

如需加入CXO UNION（CXO联盟）高管社群，请联系社群小伙伴哦~

免责声明: 本网站(http://www.cxounion.cn/)内容主要来自原创、合作媒体供稿和第三方投稿，凡在本网站出现的信息，均仅供参考。本网站将尽力确保所提供信息的准确性及可靠性，但不保证有关资料的准确性及可靠性，读者在使用前请进一步核实，并对任何自主决定的行为负责。本网站对有关资料所引致的错误、不确或遗漏，概不负任何法律责任。

本网站刊载的所有内容(包括但不仅限文字、图片、LOGO、音频、视频、软件、程序等) 版权归原作者所有。任何单位或个人认为本网站中的内容可能涉嫌侵犯其知识产权或存在不实内容时，请及时通知本站，予以删除。