In this week of advocacy on intellectual property rights, the editor has dedicated a special edition to the judicial protection of intellectual property, focusing on emerging areas and at the forefront of judicial practice, thereby demonstrating an era of judicial protection of intellectual property. The 7th edition is devoted to the theory of intellectual property protection in the emerging field of high-quality judicial protection, written by prof. Mahid, a representative of the national people's congress and a leading knowledge scholar, and explains the direction of the development of the emerging field of judicial services, along with the publication of a practical article on the determination of the liability of patent holders and post-conciliating downstream sellers, which analyses the jurisprudence of the problem of patent protection practice. The 8th edition of the study focused on trials of cases before the beijing intellectual property court and systematically demonstrated developments in the practice of specialized trials. The combination of theory and practice, macro- and micro-level, seeks to multidimensionalize the firm determination and pragmatic initiatives of the people's courts to enhance knowledge-based justice protection, service innovation and development. Attention。
Ma yidd
Representative of the fourteenth national people's congress
Professor at the university of the chinese academy of sciences and dean of the institute of intellectual property
In today's world, a new scientific and technological revolution and industrial transformation have developed, with the emergence of new technologies, such as artificial intelligence, big data, biomedicine, new energy sources, new business patterns and new patterns, driving profound changes in knowledge production patterns, industrial organization patterns and market competition patterns. Along with this, the range of objects protected by intellectual property rights, the structure of interests, tort patterns and governance logic are accelerating to reshape. In the face of profound changes in global science and technology competition patterns, building barriers to the protection of intellectual property rights in new and emerging areas has become an important basis and key underpinning for stimulating innovation throughout society and fostering new qualitative productivity. Justice plays a particularly crucial role in this process. Justice is at the forefront of the divisional debate and, in the face of emerging new types of disputes, must respond to real problems, delineate boundaries of conduct and stabilize market expectations through individual decisions. At the same time, in the absence of adequate rules in emerging areas, the judiciary has continued to accumulate experience through the administration of justice, refining rules and providing an important reference for legislative improvement and institutional innovation. Justice can be said to be both the last line of defence for intellectual property protection and an important driving force in the generation of rules in emerging areas。
Compared to the intellectual property regime developed in the industrial economy, emerging areas are more dynamic, complex and spilly in terms of technological logic, business models and equity patterns, and existing rules have in many respects revealed a contradiction of inadequate supply and poor application. Artificial intelligence depths are embedded in the whole process of knowledge production, data training and content generation are intertwined, and traditional rights systems centred on “human creation” are challenged; data, as a key production factor, have unclear property properties and access borders, constraining the development of the market for data elements; the platform's role in content generation and dissemination is growing, violations tend to be scaled up and automated, and traditional governance mechanisms are under real pressure. In this context, judicial protection of intellectual property rights in emerging areas needs to be further deepened and improved in response to practical problems. It would be interesting to explore how the boundaries between training behaviour and the right to produce content could be properly defined in the context of rapid development of artificial intelligence, how rules for data ownership that balance protection and circulation could be built in the construction of the market for data elements, and how the system of platform liability could evolve in response to abusive patterns。
I. Improved intellectual property protection systems in the area of artificial intelligence
Artificial intelligence, as a strategic technology leading to a new scientific and technological revolution and industrial change, is profoundly reshaping knowledge production patterns, industrial organization patterns and global innovation patterns. Artificial intelligence has a significant “goose effect” and is a key variable for our ability to seize the opportunities of the new technological revolution. In this context, improving the system of protection of intellectual property rights in new sectors, such as artificial intelligence, not only responds to the real needs of technological change, but is also an important support for the development of new quality productivity and the construction of modern industrial systems。
Operating logically, artificial intelligence is based on the "data input - algorithmic learning - content output" path, and its components are embedded in intellectual property elements, which raise a new set of legal issues. Traditional intellectual property systems have evolved in the age of industrialization, with “human creation” as their institutional starting point, and have gradually emerged with the application of tensions in the face of intelligent systems with autonomous learning capabilities and probabilities generation mechanisms. Particularly in the areas of data training, content generation and identification of violations, there is a growing contradiction between the supply of systems and the need for practice。
First, the boundaries of the right to conduct artificial intelligence training are not clear. Large model training, supported by big data, inevitably involves the capture, reproduction and processing of various works at the “input” stage. From a legal evaluation, these acts may involve such rights as reproduction rights, adaptation rights, compilation rights, but their main purpose is to enhance the performance of algorithms and not to directly substitute for the dissemination of original works. Strictly requiring an authorization on a case-by-case basis would be difficult to operate and might raise the innovation threshold; however, a lack of regulation could undermine the rights holder's interests and disrupt the competitive order. The question of how to strike a balance between the promotion of technological development and the protection of legitimate rights and interests is an important one to be addressed by judicial practice. Second, the right attributes of artificial intelligence-generated content need to be further clarified. The current copyright law uses “human creative contribution” as the core criterion, but the level of human participation varies significantly in the context of the generation of artificial intelligence scenarios, ranging from simple tips to complex choices and editors. As a result, there is considerable disagreement as to whether the creation of the content constitutes a work and how the attribution of rights is to be determined. A simple denial of their copyrightability may inhibit the development of new forms of creativity; and an indiscriminate empowerment may undermine the institutional foundations. A reasonable definition of the boundaries of rights in the creation of human beings in concert with the basic principles is an important direction for institutional improvement. Finally, the identification and allocation of responsibility for violations of artificial intelligence-generated content is challenging. At the “output” stage, the generation of content may be highly similar to the expression of the existing work, or even a substantial substitution, but this similarity often results from algorithmic learning and probability generation rather than direct reproduction, which makes it difficult to apply the traditional “exposure+substantive similarity” criterion. At the same time, the content is generated and disseminated mostly by the platform system, the platform is more capable of technology control and risk forecasting, the traditional user-centred “safe haven rules” are difficult to fully adapt and the platform's liability boundaries are in urgent need of adjustment。
In response to these questions, systematic thinking should be maintained and intellectual property protection rules in the area of artificial intelligence should be promoted: first, rules for training conduct should be properly defined. In the context of safeguarding the legitimate rights and interests of rights holders, the technical reproduction in model training is given a space for proportionality. Building on institutional experiences such as “text and data mining exceptions”, more inclusive regulatory arrangements for training practices that do not constitute alternative uses, based on a clear purpose, scope and conditions of use, could provide a stable source of data for artificial intelligence research and development. The second is the rational use of boundaries with clear data. Harmonizing data use and rights protection, using anti-improper competition law as the basis for a comprehensive evaluation of data acquisition and use, and recognizing legitimate and reasonable use without prejudice to competitive interests. At the same time, the promotion of orderly open public data and the enhancement of high-quality data supply capacity. The third is to improve the rules for generating content protection. In the area of copyright, the focus is on “human creative contribution”, the legal protection of results embodying substantive manual choices and arrangements and the clear identification of pathways through a typology decision. Fourth, in tort determinations, emphasis is placed on a comprehensive examination of the expression of similarities and market substitutions for new types of exploitation, such as style imitation, and on promoting the interface between copyright law and anti-improper competition law。
Overall, the impact of artificial intelligence on the intellectual property system is essentially a reshaping of the traditional “human” rights system. The dynamic balance between encouraging technological innovation and maintaining an order of rights can be achieved only by continuously refining rules and harmonizing standards through judicial practice, based on adherence to the rationale of the intellectual property system, providing solid guarantees of the rule of law for the healthy development of artificial intelligence。
Ii. Structuring a framework for the protection of property rights through anti-improper competition jurisdictions
The digital economy is becoming a key force in restructuring global factor resources, reshaping the structure of the economy and reshaping the pattern of competition, and the value of data as a core factor of production in the digital economy is increasing. At its twenty-third plenary session, the party made it clear that the establishment of a system of data titling, market transactions, distribution of interests and protection of interests should be expedited. The paper also highlighted research to construct rules for intellectual property protection of data. This suggests that the construction of a property rights protection system that is compatible with the characteristics of the data elements has become an important task in advancing the high-quality development of the digital economy。
In this context, the question of how to construct a path of protection of property rights in data that combines protection and circulation becomes a key issue for institutional integrity. The current theoretical community is more active in advocating the creation of “data property rights”, but it should be clear that data are essentially information-based and have stronger public attributes. The general exclusionary effect of granting similar property rights may unduly restrict the flow of information, raise market access thresholds and even exacerbate data monopolies. In comparative law practice, the major countries do not follow the path of creating a new type of data property right, but rather are guided by the maintenance of a competitive order and achieve limited protection by prohibiting improper acquisition and use. From the practical point of view of our country, the construction of a framework for the protection of property rights through anti-improper competition law provides greater practicality and institutional advantages, in particular the completion of the anti-improper competition law revision and the introduction of a series of guiding cases, which largely define our anti-improper competition law. At the heart of this path lies the indirect realization of data rights and interests by regulating unfair competition on the basis of the operator's legitimate control and input over the data, rather than directly granting them absolute exclusion。
In particular, attention should be paid to the improvement of judicial rules in the following areas: first, the identification of the basis of interest for data protection. Judicial decisions should identify the competitive and investment interests of enterprises in the collection of data from their inputs to data collection, collation, processing and maintenance. As long as the data are of a certain scale, stability and offer a competitive advantage to the operator, they can be the object of interest protected by anti-improper competition law. Second, the development of a “technology control plus legal protection” synergy mechanism. Under the revised anti-improper competition law data section, where enterprises have adopted technical measures to control data, there should be a clear negative evaluation of such practices as breakthroughs in technical measures and circumvention of access restrictions, gradually clarifying the criteria for judicial determination, and transforming technical controls into points of legal protection. Such a “control-based, harm-based” protection model would help to avoid excessive restrictions on the free flow of information while safeguarding the input of enterprises. Third, to refine the criteria for determining the improper acquisition and use of open data. The circumvention of technical measures to capture data, to obtain data in violation of an agreement, and to seize and make large-scale use of the data resources of others should be characterized as unfair competition, taking into account, inter alia, means of conduct, subjective malice and effects on the competitive order. Finally, the boundary space for data use is reasonably defined. Data protection does not exclude all exploitation and should be tolerated for reasonable use where the source is legitimate and does not substantially harm competitive interests. In judicial decisions, emphasis should be placed on distinguishing between “competitive utilization” and “alternative use”, which are strictly regulated by maintaining the necessary space for the former to achieve a dynamic balance between protection and circulation。
In general, the creation of a data property protection framework through anti-improper competition laws can be effective both in preventing improper seizure, guaranteeing returns on enterprise data input and avoiding the systemic risks associated with creating absolute exclusion, and in achieving a reasonable balance between protection and openness. In the context of the accelerated development of the digital economy, the role of justice in rule-making should be fully utilized to promote a clearer, stable and predictable system of rules for data protection through the continuous accumulation of adjudicative experience and refinement of standards for identification, providing solid guarantees of the rule of law for the healthy development of the market for data elements。
Iii. Towards a transformative upgrading of liability rules for intellectual property protection platforms
With the widespread application of artificial intelligence and digital technology, the patterns of intellectual property abuse are undergoing profound change, showing a tendency to replace reproduction transmission with generation, to target output from an open presence, and to accelerate the evolution of individual behaviour to scale chain operations. Unlike traditional abuses, the current abuses no longer depend on direct reproduction of existing works, but rather on “de-replicating expression” through content reorganization, style imitation, which may lead to substantial substitution of the overall effect; the content of the violations is no longer kept open, but is directed, instantaneously generated, quickly dispersed, making it significantly more difficult for rights-holders to discover and take evidence; and the violations are characterized by low-cost, high-frequency and industrialization through automatic generation and bulk distribution techniques。
