数字经济中的数据和算法规制

On the regulations of data and algorithms in digital economy

在数字市场,企业的重要策略性行为往往由数据和算法所驱动,这是数字市场与传统市场中策略性行为的重大差异。随着数据采集和处理技术的不断突破,数据体量无穷膨胀,数据采集、流转、开发、利用的场景呈现越来越强的多样性和复杂性,数据权属及利益分配很难由几部法律和文件来清晰规定。因此,迫切需要对数据和算法引入科学合理的规制。算法具有自动性、隐秘性等特点,既要考虑这些特点对规制带来的挑战,又要避免规制对创新造成不必要的伤害。欧盟在这方面的尝试值得关注,鼓励诉讼更有助于权利界定。从更宽广的视野来看,需要建立一个基于场景并嵌入未来的数权体系以及相应的算责体系,这与工业经济时代所建立的产权体系具有同样的基础性意义。

The issues of anti-trust and anti-unfair competition in digital markets are becoming a policy focus both in and outside China. This paper argues that many strategic corporate conducts, which involve monopoly and unfair competition, are driven by the use of data and algorithms, including differential pricing, self-preferencing, refusals to deal, etc. The use of data and algorithms makes those conducts automatic, secret, systematic, and changes game rules. Therefore, the regulations of data and algorithms should be the priority when the government considers how to design and improve digital market policies.But it is really hard to recognize and define the collection and use of data as well as the operation of algorithms. Data are usually not regarded as a key input of economy and not claimed to rights and benefits in traditional economy, and of course there are no techniques or rules to define data rights and to distribute data-induced benefits. Nowadays, digital economy is making a sense. In reality, more and more data and algorithm disputes are occurring and some of them are brought to government regulatory bodies and courts. These scenarios urge some countries to establish regulations. The European Union Parliament has passed General Data Protection Regulation(GDPR) and Data Governance Act. Some House members in the United States have put forth proposals on algorithmic accountability act. China is also pushing ahead in this field. But it is fair to say all these pilot regulation trials are initial actions, which cannot cover most of the actual issues of data and algorithms, and cannot deal with the expansion and development of data and algorithm utilization as well. The analysis of this paper points to the context-based means so as to manage data and algorithm issues in digital economy. The utilization and profiting of data and algorithms involve unnumbered contexts, and specific definition of rights, benefits and responsibilities will be given in each context. Therefore, context-based judgments are able to respect accurate facts and intentions, to supply more clear and correct remedies, and of course to provide more incentives for data altruism. That means the regulatory pattern should resort both to government agencies and courts. The legislation can determine major principles and frameworks, and it is a good idea to set up professional data courts to explore and develop specific definitions and discretions. This system can tackle complex cases and build up step-by-step rules, as well as train new judges and practitioners.China faces some distinctive issues when regulating data and algorithms. China is a socialist country and has a different perspective of state security and economic security. Therefore, it lays more emphasis on Internet sovereignty and data security. Actually, many countries are alert to those concerns, and it is understandable that China is more alert. China has published some legislations and regulations to strengthen data and algorithm security. As data need to have cross-border circulations and uses, China will face the question of how to avoid data islanding effect.