Sun Shan: On the Legal Status of Humanoid Robots

Sun Shan is an associate professor at the School of Civil and Commercial Law, Southwest University of Political Science and Law.

Robots that integrate artificial intelligence, big data, and algorithms are undoubtedly the next industry in the wind. Among the many types of robots, humanoid robots are the most striking and thought-provoking. Humanoid robots not only have the common characteristics of robots, but also have technical, ethical and legal issues that require special attention due to their human-like appearance and deep involvement in human family life. The embodied existence of humanoid forms and the decisions and contents generated in an emergent way, the innate positioning of tools and the acquired projection of human emotions all push us to think about the meta-problem of humanoid robot regulation: the legal status of humanoid robots. The three options in front of humans are: keeping up with technological changes and directly recognizing the dominant position of humanoid robots; Stick to the human standard and adhere to the object positioning as a tool; Adopt a compromise attitude and recognize its dominant position to a limited extent. After completing the initial selection, we still need to achieve the corresponding legal status at the institutional level, sublimate the existing legal theory and legal normative system, make it more fair, effective and flexible, and keep up with the rapid development of the digital age.

Limited by the development level and years of the industry itself, there are few studies on the legal status and realization of humanoid robots at home and abroad. At present and in the foreseeable future, artificial intelligence is the basic technical framework for the operation of robots, and some scholars do not strictly distinguish between artificial intelligence and robots, robots can be regarded as the embodiment of artificial intelligence, and the current discussions on artificial intelligence legal issues are also applicable to all types of robots, including humanoid robots. The embodied existence of humanoid shape is the special feature of humanoid robots that distinguishes them from traditional robots, and the emergence of decision-making and content generation methods is the practical basis for the relationship between humanoid robots and humans. Starting from the problems raised by embodiment and emergence, this paper explores the considerations behind the changes of civil subjects by combing the history and current situation of robot development, and arguing the core view that humanoid robots can only be identified as legal objects rather than legal subjects from an anthropocentric standpoint.

1. Embodiment and emergence: Interpretation of the phenomenon of humanoid robot operation

There are many legal issues caused by the application of technology in the operation of humanoid robots, and the root of these legal problems is embodiment and emergence. American scholar Ryan Calo believes that there are significant differences between robotics and the Internet, which have led to a series of new legal issues, which are mainly reflected in the following: for the first time, robotics combines the promiscibility of data and the ability to cause physical harm through the way of embodiment; Robotic systems complete tasks in unpredictable ways, i.e., emergent ways; Robots blur the lines between humans and tools. Embodiment and emergence are the factual basis for various legal problems caused by humanoid robots, and naturally become phenomena that we must be cautious when establishing the legal status of humanoid robots.

(1) Embodiment: The starting point for humanoid robots to deeply intervene in human life

Embodimentation in a human-like shape is the most obvious difference between humanoid robots and ordinary robots. First of all, it should be stated that the robots studied in this article are robots that exist in the physical world in an embodied way, excluding "software simulator robots" that exist in computers. The so-called embodiment refers to the mechanical device that takes all kinds of mechanical devices as the entity of its existence and concretizes various program functions into various executable instructions. Humanoid robots, on the other hand, refer to robots that are close to human appearance in terms of overall shape. Robots have been widely used in human social life and family life, and there are various forms of existence, but most robots have not appeared in a humanoid way. In fact, humanoid shapes will increase the cost of design and manufacturing, and only those robots used to handle care and intimate affairs need to be designed into human-like shapes to increase human acceptance of robots. Industrial robots, special robots, etc. do not need to be designed into human-like shapes at all, and the legal issues raised by these two types of robots will not be discussed in this article. In terms of overall contour and shape, we can still distinguish humanoid robots from ordinary robots, and the human-like shape is the most obvious difference from ordinary robots, and this difference raises a series of legal issues.

Embodiment is the beginning of the deep involvement of robots in human life. Different from the technological inventions used by human beings in the past, robots are more invasive, sensors, network interconnection and algorithms are the technical basis for robots to invade human life, and we are facing the risk of being spied on, recorded and imaged at all times. First, the robot will actively collect and transmit all kinds of information. In the past, various technological inventions did not have the ability to actively collect information and transmit it to the outside world, but were just simple tools. The "invasiveness" of humanoid robots is mainly reflected in the active collection and transmission of information, and private houses have also become the scene of big data acquisition, and the connection with the Internet makes it possible for information to be continuously transmitted to the cloud, and private information has the danger of becoming public information. More than that, due to the gradual dilution of privacy boundaries, the free space of individuals is further compressed. Second, the robot integrates a number of core technologies, which can complete the portrait of the user, and amplify the invasiveness in disguise. With the blessing of Internet technology, mobile terminals such as platform computers and smartphones can also collect and transmit information, invading human social and family life. The application of algorithms empowers platform operators, who can extract useful information from the massive data obtained, but these massive data do not contain specific information about the network user's home life. The embodied presence of robots is more invasive than any tablet, smartphone, or sensor, and with the help of algorithms, the user's living habits, personality traits, personal preferences, etc., are also known by the manufacturer. The more powerful the feature, the greater the threat.

Embodiment enables robots not only to enter human family life, but also to affect human social life. Robots have permeated every aspect of human life, and this change has forced us to face the social impact of robots. The main impacts of embodied robots on human life are: first, physical damage at the physical level. The embodied physical presence inevitably poses harm to humans and the associated problems of responsibility, and humanoid robots that walk into homes extend harm from factories to homes. Second, the psychological level of privacy destruction. Normally, humanoid robots can freely enter and exit every room in the home. In the absence of legal regulations, humanoid robots that walk into the home will continue to collect users' private information, which will eventually have a devastating impact on privacy. Third, there are multiple threats to public security in the data field. If there are no restrictions on the information collection by bots, then it is conceivable that there will be other entities outside of the platform who misuse other people's private information, including but not limited to businesses, individuals, and governments. When each of us becomes transparent because of the disclosure of privacy, the private sphere constructed with the boundary of the house will no longer exist, and the public safety of the data field will exist in name only.

The embodiment of humanoid robots further stimulates the imagination of human beings to recognize the qualifications of humanoid robots. The similarity in appearance will make us empathize, and empathy is the moral starting point of the humanoid robot subject qualification imagination. Some foreign scholars believe that compared with any previous technology, the similarity in appearance makes the robot have a greater degree of social utility to humans, and the status of the robot is between people and objects. Some people are more inclined to see humanoid robots as the next subject to be treated as equals than some domestic animals that have been with them for a longer period of time. Although some domestic animals are emotionally more closely related to humans, there has been no systematic legislation in the world that treats other animals as legal subjects for special protection, at most emphasizing that other animals are special objects of rights, and rejecting acts that violate human moral consensus, such as torture and killing. At the end of the day, it's the emotional need to protect other animals. The humanoid embodiment and communication ability gives us the urge to see humanoid robots as our kind, and this urge gradually evolves into a continuous appeal as technology advances.

(2) Emergence: an alternative logic formed by humanoid robot decision-making

The emergence is the logic formed by the corresponding decision-making of humanoid robots using artificial intelligence technology. emergence, which is a translation of the English word "emergence". Emergence, as a scientific term, was first used in 1972 by the famous physicist Philip Anderson, who described emergence, as an ability that does not exist in a smaller model but occurs in a larger model. The emergence reveals to us the effect that when the number of individuals reaches a certain size, a group phenomenon cannot be deduced from the original individual phenomenon, and that the behavior that complex physical systems may exhibit cannot be understood solely by the laws governing their microscopic components. Emergence is a common phenomenon in the application of artificial intelligence, especially large models. With built-in artificial intelligence, humanoid robots can also emerge during operation. From the very beginning, robots have been able to complete certain tasks according to the instructions given by the program. However, for a long time, robots can only strictly perform assigned tasks, and when the external environment changes, robots cannot adapt. With the rapid development of supporting technologies such as artificial intelligence and big data, robots have gradually acquired a certain adaptability, and can randomly take "autonomous action" according to the environment.

The responsibility for the damage caused by the emergence is more complex, and the issue of attribution has become a prerequisite that must be clarified in the process of commercial application of humanoid robots. In the early stages of development, robots could only execute pre-given procedural instructions by humans, and could not change their behavior according to the external environment. In this technological context, there is no doubt that the damage caused by the robot is the responsibility of the person who gave the instruction. However, because it cannot respond to changes in the external environment, the robot at this time can only complete specific task types and is a dedicated robot. As we all know, special robots are more suitable for industrial assembly lines, and the cost of entering family life is high, and general-purpose robots are a rational choice for family use. Artificial intelligence, especially the application of large models, empowers humanoid robots to interact with the external environment, which greatly promotes the development process of humanoid robots, and general-purpose robots have the realistic possibility of landing. One of the fundamental differences between robots and computers is their interaction with the dynamic world. "The robot is constantly interacting with the dynamic world, and often the algorithms set up are not enough to cover all of these dynamics, so the final behavior of the robot is very different from what is coded." The final behavior transcends the encoded content, that is, the emergence of artificial intelligence, big data and algorithms. Emergence means that when humanoid robots interact with the external environment, humans are not able to accurately predict the direction and outcome of their behavior. As a result, it's not clear who is responsible for what the code does.

The attribution of harm caused by humanoid robots should be free from the influence of opportunistic thinking, and the principle of consistency of "rights-obligations-responsibilities" should be emphasized. In his view, code rooted in software and hardware defines the terminology of cyberspace, becomes the regulator of cyberspace, and belongs to the "preset environment" and "architecture" of social life, which can constrain and guide human behavior. And when emergence, the opposite is true: self-learning systems may be neither predictable nor subject to human expectations of appropriate behavior, and code becomes a tool to evade human regulation and is no longer the law that regulates humans. On this basis, some scholars further pointed out that opportunistic thinking often appears in the discussion of robot law, and some people will dilute and emerge when they benefit from robots, and some people will emphasize the emergence when robots cause harm, and the goal of the law is to satisfy this foreseeable opportunism and improve the distribution of benefits and the attribution of responsibility. In this regard, this paper holds the opposite view, and the attribution and empowerment of humanoid robots should adhere to the principle of "rights-obligations-responsibilities", and the risks and benefits should be integrated, so as to construct an institutional system that conforms to the values of fairness, efficiency, safety, and order. In the process of market promotion of humanoid robots, the design and manufacturer are the main beneficiaries, and the purchaser can also benefit from the subsequent use behavior, and the performance of obligations and the assumption of responsibilities should also be mainly distributed among these parties.

The consistency of "rights-obligations-responsibilities" is also the leading principle to solve the problem of whether the content generated by humanoid robots is protected by the Copyright Law. At the same time, humanoid robots will also bring valuable output to users - AI-generated content. Due to the built-in artificial intelligence and the embodiment and externalization of mechanical devices, humanoid robots will not only generate content like ordinary artificial intelligence, but also output and fix the generated content into various physical forms that can be perceived by humans due to the existence of the embodiment, and also face the trade-off of whether to protect copyright. From the perspective of the generation process, the emergence is inseparable from the participation of human subjects. The premise of the emergence is that a human as a user enters a basic instruction, and the AI then generates content that conforms to the basic instruction. If they need content that is more relevant to their requirements, users will also have to enter more qualifying conditions to interact with the AI. Obviously, the more restrictive conditions a user enters, the more content is generated that meets his needs. From the point of view of the results generated, the resulting results are indistinguishable in form from the intellectual achievements created entirely by humans and done independently. Originality is the constituent element of a work, and it is the key to determining whether a particular expression constitutes a work. In the final analysis, "the object of judging originality is the expression itself, which has already been formed, and only the formal examination of the expression is required to be judged." Some scholars advocate the exclusion of the originality of AI-generated content, but if the subject of content generation is not informed in advance, it is impossible to effectively distinguish between human beings alone, let alone complete legislative and judicial practice on this basis. Acknowledging the originality of emergent results is to transcend human bounded rationality on the basis of acknowledging human bounded rationality, and ultimately serve rational human beings. If the rights, obligations and responsible subjects are different, the legal and commercial risks of content generation will be infinitely magnified, and the development of the industry is bound to become empty talk. Emergence, by no means repetition, is a cheaper and more effective solution to the problem of generative AI by treating it as a creation based on the principle of "rights-obligations-responsibilities", and identifying the subject of rights as the user according to the principle of "rights-obligations-responsibilities".

2. Appearance: The negation of the legal subject status of robots

(1) The academic entanglement between the subject of rights and the object

In traditional narrative expressions, robots are often defined as a special tool that has the potential to eventually rebel and enslave humans. In literature, robots are often described as being difficult to control due to a lack of complete programming or a sudden occurrence of behavior or error. In film and television dramas, robots are the hardest hit area of demonization, "Star Wars", "Blade Runner", "Terminator" and "Westworld" have all launched impressive robot villains, and the survival of human beings themselves is threatened by robots. It is undeniable that people in contemporary society are greatly influenced by the media, and the image of mechanical, numb, and cold-blooded robots created in traditional narrative expressions will affect our cognition and acceptance of humanoid robots themselves. The famous science fiction writer Asimov played an important role in the reversal of the image of robots: "By creating the Three Laws of Robots, he transformed our imagination of robots from a threat to an object of interpretation and regulation, and thus a source of irony and conflict." ”

With the rapid development of artificial intelligence technology and the increasingly serious aging population, confirming the dominant position of robots has become a relatively fashionable discourse. According to the scope of application of robots, they can be divided into industrial robots, service robots and special robots. It is foreseeable that in the context of the aging of the global population, service robots will enter the daily life of ordinary people. In this process, we will inevitably think about the same problem faced by different countries: whether to confirm the dominant position of robots. For industrial robots, we usually do not have the idea of giving them a dominant status, and the two characteristics of being different from human appearance and loosely related to human family life make it difficult for humans to have emotional projection on them. Service robots, on the other hand, are closely related to human family life and cause human emotional projection. Humanoid robots inspire human empathy because of their shape, "sensitive ethical issues arise when robots evoke trust or friendship between vulnerable groups", and the core of ethical issues is whether to establish the subject qualifications of robots and treat robots equally. Some scholars even imagine the symbiosis and integration of "cyborg" culture and human culture in a human-machine hybrid society, accept the legal subject status of "cyborg", and choose the basis of their argument in the distant and unknowable future.

The characterization of the subject of rights and the object of rights will have different legal effects, which is the institutional background that we must consider in the process of clarifying the legal status of robots. The fundamental difference between the subject of rights is the end itself, and the object of rights is the tool to achieve the goal, which determines that the object of rights is a category that can be disposed of arbitrarily, and the subject of rights has its own independent value. Whether it is the subject of rights or the object of rights, as functional concepts, their value is dependent on a specific normative system, and the difference in qualitative means that the legal effect is very different, which means that the supporting norms are obviously different. This reminds us that when clarifying the legal status of robots, we must not only consider the philosophical and ethical preferences of individuals, but also take into account the legal and social effects of legal qualitative behavior.

(2) Considerations behind the change of civil entities

The confirmation of the qualifications of civil subjects is the result of social development to a certain stage, and there is no a priori and unchanging scope of civil subjects, and the need for economic exchanges is the primary factor considered by legislators. The subject of law has undergone a historical change from "people can be non-human" to "people can be people" and then to "non-human people". From Roman law to the legislation of modern countries, the equal status of natural persons has been gradually recognized, and the ranks of "humans" have been continuously expanded. This expansion is due, on the one hand, to the flourishing of the concept of human rights, and on the other hand, to the disintegration of the extended family, the independence of the small family and the change of the individual as a unit of economic activity, as a result of economic development. At the same time, the status of legal persons and unincorporated organizations has also been recognized by legislators, and the recognition of these two types of subject qualifications has reduced the unlimited liability of natural persons to a certain extent, stimulated economic vitality, and performed public functions that individuals cannot perform on their behalf. Therefore, the subjectivity of non-human entities is derived from the abstraction and virtuality of the legislator, based on the needs of real people. From the historical combing, we can draw a conclusion: the primary consideration factor for the granting of subject qualifications is the need for economic exchanges, followed by the requirements of ethics and morality, and the core issue of the subject system is always property, and who owns the property.

The main purpose of the legislator's confirmation of the qualification of civil subjects is to clarify the ownership of rights, and the premise of recognition is that civil subjects can effectively communicate and empathize with each other, and ethical care is a secondary consideration when confirming the qualification of civil subjects. Whether it is a natural person, a legal person, or an unincorporated organization, the mainland civil code confirms its subject status in order to complete the subsequent empowerment. In addition, civil subject qualification also has its ethical value. The example in this regard is the protection of the interests of the deceased and the fetus, while the counterexample is the denial of the qualifications of other animal subjects. The protection of the interests of the deceased extends the legal remedy to the death, and the protection of the interests of the fetus extends the legal remedy to the time before the birth. The extension of the two directions is not based on the definition of "person" in everyday language as a legal premise, but out of the need to protect the ownership of rights and realize ethical care. The deceased once enjoyed the right to perform the obligation, and the fetus will have the right to perform the obligation in the future, and the social relationship with the content of "right-obligation" is the essential characteristic of them as a social person. Ethical care, on the other hand, is to shape a complete "person" at the moral level. In contrast, other animals with a certain amount of intelligence, emotions, and some human emotions are completely excluded from the scope of subjectivity, and the laws made by humans have always been aimed at safeguarding human interests. We cannot know for sure the emotions of other animals, let alone their true intentions, and there is no feasibility and necessity for further implementation of corresponding legal acts. In a situation where other animals can never exercise their rights and fulfill their obligations independently, giving them "rights" naturally has no value at all.

The confirmation of the status of civil subjects is based on the assumption of civil liability, and the construction of a free, fair, just, safe, efficient, and orderly legal system environment. Civil subjects must not only have the capacity for rights, but also have the capacity to act, and be able to bear responsibility for their illegal acts, and the capacity for responsibility is the proper content of the capacity for conduct. Freedom, fairness, justice, security, efficiency and order are the basic legal values pursued by human beings, and the realization of these legal values is inseparable from the responsibility system. The freedom to be able to take responsibility for one's own choices is worth pursuing, and the freedom to choose not to be responsible can be a disaster for others. Fairness and justice can only be upheld by taking responsibility for the damage caused to others. Taking responsibility is the bottom line requirement for the safety of the people, and the subject of the law has a minimum stable expectation of the behavior of others. The attempt to assume responsibility in exchange for the pursuit of efficiency is a common behavior in the field of civil activities, and efficiency breach of contract is also a rational embodiment. By pursuing the responsibility of the subject of breach of contract and violating the law, and giving negative normative evaluation, the liability system constructs a stable legal order. To sum up, the assumption of responsibility ensures that the confirmation of the status of the civil subject tends to be consistent with the value pursuit of the law, the legislative purpose is realized, and the independent assumption of civil liability becomes an important part of the qualification of the civil subject.

(3) Denial of the qualifications of the robot entity

The characteristics of the robot at the physical level determine that it cannot understand the meaning of legal norms for itself, and legal norms cannot directly affect the robot's choice of action, so it is meaningless to formulate that the legal subject must have the right capacity for the robot. First, robots can understand the literal meaning of legal norms, but they can't really understand the actual effect of legal norms. Robots have no need to obtain property, no expectation of forming a family and extending life, and cannot feel the finite nature of life and the preciousness of freedom. "Right capacity is the ability to be the subject of rights and obligations", "rights-obligations" is the genealogy of human natural needs and social relations, is the norm of legal relations, all the "rights-obligations" as the form of expression, based on people's natural needs and social relations of legal norms, it is impossible for robots to recognize them from the level of practical effectiveness. Second, robots have no incentive to acquire property and other social evaluations, and their intrinsic driving force is computer code. Robots do not have the interests that must be satisfied through monetary transactions or social evaluation, and they cannot understand the essence of the entire legal system, including the property system, and giving them a subjective status is not conducive to the effective operation of the legal system. Regardless of whether there is a transfer of interests, the robot will carry out the corresponding behavior, and the built-in program code is its driving force, and it cannot make behavioral choices autonomously and based on its own interests. When legal norms can neither be truly understood by robots, nor can they directly and effectively regulate the process and results of their action choices, the confirmation of their legal subject status will naturally lose their legal significance.

Robots cannot take responsibility independently and do not have the responsibility capacity required for legal subject qualification. At present, most of the scholars who advocate the subjective status of robots avoid the specific realization of responsibility, and a few suggestions on civil liability are stacked on the bed frame, advocating that in the case of robots as the main body, humans should spread the risk by means of commercial insurance, the establishment of compensation funds, etc., the essence of which is that humans rather than robots bear responsibility. This way of bearing civil liability by not admitting half a pot of wine money when drunk to death just denies the legal subject qualification of robots. The assumption of criminal liability is similar to that of civil liability. Robots do not have life in the biological sense and cannot truly understand the limits of life and freedom. Some scholars believe that when an intelligent robot that has reached a certain stage of development commits a criminal act in accordance with its autonomous consciousness and will outside the scope of program design and preparation, it is entirely possible to become the subject of the actor and bear criminal responsibility. In their view, intelligent robots have autonomous consciousness and will, and can perceive the "pain" of punishment, and then seek advantages and avoid disadvantages to complete behavioral choices. But for a being who does not need property and cannot truly understand life and freedom, the execution of any punishment is meaningless. To sum up, robots do not have the capacity to be responsible, and forcibly stipulating them as legal subjects will only be counterproductive.

It is difficult to reach an ethical consensus in a short period of time for the establishment of the subject qualification of robots, and the legitimacy is insufficient. First, humans are unable to empathize with robots in the true sense of the word. The human-like shape and artificial intelligence technology can give some people the illusion that humanoid robots can communicate with humans without restrictions. However, the real situation is not developing as fast as imagined, robot technology is still in its infancy, humans and robots lack the basis for mutual empathy, and we do not need to amplify the unilateral emotional projection of some people into the emotional exchange between the two sides. Robots, in turn, are unable to empathize with humans. The existence and growth of each person's life is a social, historical, and cultural process, and correspondingly, human beings have a triple life of physical life, spiritual life and social life, and social life may be a technical bottleneck and value paradox that robots cannot surpass. Second, the purpose of recognizing the qualification of robot subjects is to reflect the ethical concern of human beings, but the result will bring more ethical problems to human beings, and the result is completely contrary to the purpose, which is undesirable. If the subject qualification of the robot is recognized, the manufacturer and user of the robot can be exempted from the attribution of responsibility caused by the damage caused by the robot, and it is natural to respect the subject qualification of the robot itself. However, "the law is inherently a precautionary measure...... There is no point in trying to apply our traditional notions of punishment directly to robots." When the result of acknowledging the qualification of robot subjects is to make human beings as moral agents evade the responsibilities that should be borne by them, this kind of institutional design that constantly lowers the bottom line of human morality should not be the direction we pursue. Third, the attempts of individual countries to confirm the qualifications of robots or natural things have limited normative significance, no universal possibility, and no reference value. Some scholars have taken the Paro robot as an example of completing the household registration in Japan, and proposed that the robot has gone beyond the "property" positioning. Some scholars have also argued that the humanoid robot "Sophia" was granted citizenship by Saudi Arabia as an argument to prove the subject qualification of the robot. However, the above incidents are extreme cases, and the symbolic value is far greater than the actual value, which is limited to after-dinner conversations.

The existence of legal fiction technology and the legislative practice of legal persons constituting legal subjects cannot directly deduce that robots should also be simulated as legal subjects, and there is no necessary connection between the two, and the simulation of robots cannot achieve the legislative purpose. Legal fiction is the theoretical support for some scholars to advocate the legal subject status of robots, and it is the legislative technical preparation to prove the subject status. It is generally accepted that legal fiction is "an irrefutable and decisive fiction made by the legislator in order to achieve the institutional purpose behind the law". The formulation of norms is to facilitate the interaction between people, and the emergence of various fictitious subjects with legal persons as typical is also to achieve this purpose, and there is no value independent of human beings themselves, "the legal person system separates the scope of rights of the group itself from the scope of the rights of its members, so that legal relations can occur between the whole and individual members". The emergence of robots is no exception, in the prospect of its relationship with people should not fall into the "value blind spot" of technocratic linear thinking, how to treat people after the development of robots, in fact, how people understand and treat people. At present, robots do not have independent interests and needs, and it is difficult to effectively communicate with humans on all matters, let alone empathy with each other, and it is impossible to achieve the legislative purpose by impersonating them as civil subjects. Through the phenomenon, some scholars pointed out that the real harm of robots to the fate of human populations is not the unwarranted robot "personality" itself, but the very small number of people in the industry who create algorithms and drive the operation of society through artificial intelligence use the cover of robot "personality" to complete the algorithmic dictatorship and technological kidnapping of more people. The development of robots has never been for the purpose of eliminating human beings themselves, but for the sake of a better life for human beings, which is our position to clarify the legal status of robots. Between fiction and imagination, we must choose reality, choose all this that we have to face - the robot is not the subject.

The existing normative system can solve the problem of creating a robot subject, but it lacks the necessity of confirming the subject. According to the Occam's razor principle of "if it is not necessary, do not add entities", if the relevant problems can be solved under the existing normative system, then there is no need to start from scratch, establish the main position of the robot, and design a complete set of normative systems based on it. The main issues to be responded to by the creation of the robot are the ownership of the copyright of the emergent results and the attribution of liability for infringement caused by the embodied operation. For these two problems, the cost of solving these two problems is obviously lower by assigning the rights to the user and the responsibility to the user or the manufacturer according to the specific circumstances, and basically following the existing normative system. There is a view that when there is already a creation or damage caused by one's own decision, AI should be regarded as a civil subject, but after being regarded as a subject, the provisions on product liability or damage caused by animals cannot be used to clarify the attribution of liability, and the provisions on employer's liability can be applied by analogy under the current legal framework. Applying the employer's liability by analogy in the absence of an employment relationship, ignoring the product liability scheme formed by positioning the robot as a tool, is not preferred. So far, there is no problem that can only be solved by establishing the legal subject status of robots.

3. Special objects that require special regulation: the legal status of humanoid robots

(1) Objects with overall controllable legal risks: humanoid robots as special objects

The fact that humanoid robots have shown a higher level of intelligence than humans in some fields and even in all fields does not automatically prove the legal subject status of humanoid robots, and the intelligence level is not a constituent element of legal subjects. We cannot give humanoid robots the status of legal subjects just because their intelligence level and overall ability to solve problems are higher than that of humans, and safe, efficient and reliable robot "tools" are the ideal image of robots. Nor do we confer the status of legal subjects on other animals, especially primates, because of the level of intelligence they exhibit. Historically, intelligence has always been only one of the legislators' excuses in the arguments for the denial of slaves, domestic children, and the status of female subjects, but this excuse is in fact not valid at all, and the purpose is to limit the ability to independently acquire and dispose of property. In the process of expanding the scope of natural persons, the substantive reason is to independently acquire and dispose of property.

From the perspective of operational effect, the system of right ownership, right exercise and responsibility attribution established by positioning the humanoid robot as the object is the most concise. The legal consequences of the humanoid robot's implementation of the corresponding behavior, including the ownership, exercise and distribution of responsibilities arising from the behavior, are reduced to tools by the positioning of the object, eliminating all kinds of troubles when locating the subject. First, after positioning as the object, the rights of the content generated by the humanoid robot belong to the user, and the subject of rights is clear, which can promote the healthy development of related industries. Second, after positioning as the object, the user actively uses the humanoid robot by virtue of its dominant position, and makes corresponding behavioral choices by seeking advantages and avoiding disadvantages. Third, after positioning as the target, the user and the manufacturer of the humanoid robot will be held responsible according to the specific situation, and while the third party receives relief, the user of the humanoid robot will also adjust the behavior strategy, and the manufacturer will improve the design. The essence of positioning humanoid robots as objects is to restore all legal effects to people. There is a view that AI has a limited legal personality and cannot be independently and fully assumed, and should be borne by its designer, developer, manufacturer or user according to the actual situation. However, artificial intelligence is not "unable to take full responsibility independently", but "completely unable to take responsibility", and behind the rhetoric of "limited legal personality" is the actual effect of the tool. No matter how the wording is adjusted, as long as the responsibility ultimately falls on humans, humanoid robots cannot become legal subjects.

At present and for a long time to come, humanoid robots can only be objects in legal relationships. Ethics and law are not only literal, but more importantly, the interaction between people that is deeply embedded in the social background behind the words. Robots do not have their own interests and do not have the internal motivation to drive them to establish social relations, and they will not form social relations with independent significance. Therefore, "at least for the foreseeable future, robots will not fully understand human morality and law". What deserves the attention of people in the legal profession is not the details of the technical implementation, let alone the possible direction of technological development, but the practical consequences of the implementation of the technology. The biggest characteristic of "technocratic linear thinking" is that it is superstitious about technology, ignores the value premise of technology itself, departs from the scope of human concern and consideration, and enters the "value blind spot". Robots are born for people, which is a cognitive stance that breaks the "technocratic linear thinking". We must ensure that the existence of humanoid robots does not endanger the survival and development of humans themselves, and this basic requirement can only be achieved when humanoid robots are positioned as objects.

Only by positioning humanoid robots as objects of rights and adhering to the position of anthropocentrism can it be possible to construct a legal order that conforms to the interests of human beings. The ancient Greek philosopher Protagoras once said that "man is the measure of all things", positioning human beings as the center of observation of things, and starting the process of evaluating the whole world with human beings as the center. It is the basic consensus of human society that law is formulated to protect the interests of human beings, and law itself is a product of anthropocentrism and belongs to the category of instrumentality. However, there are always some scholars who have a different starting point, thinking from the perspective of the so-called God, and using the narrative of fraternity and pan-love that overthrows anthropocentrism to understand and transform the world, and the legal status of robots is no exception. Some scholars in the United States put forward the theory of "strength defines rights", emphasizing that the power game of different groups determines the construction of the rights system. Some scholars in China have taken this step further, proposing that the generation of robot rights completely depends on the strength comparison and game between robots and humans, and that a legal guarantee mechanism for the coexistence of different races should be established to cope with the huge pressure of an aging society. According to this logic, humanity will eventually fall into a very absurd paradox: if humans continue to advance the research of robots, then robots will gain dominance over humans because of their increasingly dominant position, and humans can only become slaves of robots; If human beings stop researching robots at some stage, then this deliberate containment will also infringe on the basic rights of robots, just as some humans curb the development rights of other people of the same kind, and the subject of rights becomes a play on words. Anthropocentrism is based on the long-term and common interests of human beings, and it is completely wrong to interpret anthropocentrism as a dwarfing and one-sided interpretation of not respecting nature and other life forms, and respecting nature and other life forms also belongs to the long-term and common interests of human beings. The anthropocentric stance is the fundamental starting point for us to build, develop and improve the system of legal norms.

As a product of human construction, humanoid robots are generally controllable objects of legal risk. In an age of lack of self-awareness, humanoid robots are not scary, and what we should be worried about is that increasingly powerful humanoid robots are not at the same pace as human goals. The human-like form factor and the integration of highly complex technology make humanoid robots hide certain legal risks. As runtimes increasingly rely on artificial intelligence, algorithms, and big data, humanoid robots will also form their own targets. In order to ensure the same goal, we need to incorporate ethical rules at the beginning of the design, introduce human value judgments, and limit the goals of humanoid robots to a wide range that is consistent with the basic consensus of humans. As for Harari's prediction in A Brief History of the Future, in which artificial intelligence will become dominant and a small number of "godmen" evolved from Homo sapiens can be "left alone", we are trying to avoid it. After the completion of the first nuclear test, the physicist Oppenheimer expressed his ambivalence by quoting a line from the ancient Indian epic poem Bhagavad Gita: "Now I am the god of death, the destroyer of the world." "If the development of humanoid robot technology is allowed to be allowed to confirm its status as a legal subject, then we will also express the same lament in the future. The positioning of humanoid robots as objects is to control the risks of humanoid robots and ensure that the legal risks are generally controllable. When humans talk about robots, they usually focus on whether they create problems or pose a threat, but in most cases, the problem is not with the robots, but with the humans themselves. In the final analysis, the various norms formulated around the legal risks of humanoid robots are used to adjust the relationship between people, rather than the relationship between humans and humanoid robots.

(2) Rights objects with higher ethical and technical risks: humanoid robots that require special regulation

Ubiquitous sensors and cloud storage threaten human privacy at all times, and human robots that are deeply involved in human life must be specially regulated. Humans are introducing service robots into home life, which may further reduce the chances of privacy protection. As some scholars have noted: "A personal bot will know everything about you, the most intimate parts of your personal life, your family, finances, and physical history." More critically, they are connected to the Internet. "Sensors actively collect and transmit information, while cloud storage retains private information permanently, amplifying the time and space of damage. For the sake of profit, manufacturers will seek excessive acquisition of personal information through humanoid robots, and we have fully experienced the troubles caused by excessive acquisition and leakage of information in the use of smartphones. Not only that, but manufacturers also trade data, selling vast amounts of data containing private information to third parties without the individual's knowledge. We must be aware that surveillance often brings immediate, concrete benefits, but also longer-term, more decentralized dangers.

The decision-making and management activities of humanoid robots that deviate from value judgments will cause harm to human beings that are in line with individual rationality but not in line with the overall value pursuit. Under normal circumstances, humanoid robots that have been rigorously designed are tools that are safe to use and help humans achieve a variety of legitimate purposes. However, when the user issues malicious instructions, because the humanoid robot lacks sufficient information to complete the value judgment in a timely manner in the specific scene, the corresponding decision-making and management activities will cause harm to human beings that is in line with individual rationality but not in line with the overall value pursuit. We can build various ethical rules into humanoid robots in the form of code, but the implementation of these abstract ethical rules also needs to be combined with specific scenarios, and the information in the scenes is often not fully known by humanoid robots. For example, if a user issues instructions to a humanoid robot to carry things, before everything is connected to the Internet of Things, the information asymmetry makes it impossible for the humanoid robot to distinguish the ownership status of the things it carries, and the user's stealing instructions can only be followed. Therefore, it is necessary to clarify at the legal level the attribution of responsibility when the user gives instructions in bad faith.

The data security of humanoid robots in operation also needs to be paid attention to, especially hacker attacks. Information leakage often occurs during data transmission, mainly due to network problems and hacker attacks. In the age of big data, there is almost no so-called "useless" data. Even seemingly innocuous information can be used to collect our personal information, and if it cannot be used alone, it can be combined with other data. During the operation of humanoid robots, hackers will threaten data security in the following three aspects: First, hackers use technical means to illegally obtain all kinds of privacy-related data. Data, whether stored or transmitted, can be stolen by hacking. Second, hackers remotely manipulate audio-visual sensors to obtain hidden video images. The use of audiovisual sensors provides a convenient way for hackers to directly peek into the private lives of individuals, and extortion ensues. Third, hackers remotely manipulate humanoid robots to carry out sabotage. Virtual attacks are moving from imagination to reality, and the remote control of humanoid service robots will cause us direct physical damage. How to prevent hacker attacks will be a difficult problem that humanoid robot manufacturers must be cautious about.

Obtaining legal authorization or complying with legal rights restrictions is a legal guarantee for the content generation of humanoid robots. Since humanoid robots use artificial intelligence as the underlying technology, the problems and countermeasures in the adjustment of the copyright law for AI-generated content are also applicable to humanoid robot-generated content. For the legal regulation of generative AI, there is a distinction between the front-end and the back-end. We usually focus on the back-end issue, which is the copyright protection of the content itself generated at the output stage. The problem of the front-end is the premise for the realization of content generation at the input end stage - the legitimacy of data acquisition, that is, the choice and application of the system of negotiation permission and rights limitation, including fair use rules. In this regard, there are those who advocate the reform of the fair use system, those who advocate the legality of machine learning behavior based on the distinction between "expressive use" and "non-expressive use", and those who advocate the use of statutory licensing to balance the interests of all parties in the industrial chain, but no basic consensus has been reached. Article 7 of the Interim Measures for the Administration of Generative AI Services stipulates that generative AI service providers shall "use data and underlying models with legal sources" and "shall not infringe on the intellectual property rights enjoyed by others in accordance with the law if intellectual property rights are involved". The above-mentioned provisions provide a framework for the legality of data acquisition, and how to implement it still needs further research.

Under the superposition of the complex technical composition of humanoid robots, acting according to instructions, and the randomness of emergence, it becomes more difficult to attribute the liability for the infringement of humanoid robots. The various drawbacks inherent in the algorithms that support the normal operation of humanoid robots need to be eliminated at the legislative level. We are moving from an internet society to an algorithmic society. The so-called algorithmic society refers to "a society organized around social and economic decision-making by algorithms, robots, and artificial intelligence agents", and robots and artificial intelligence are one of the application cases of algorithmic society. However, the algorithmic black box makes it possible for humans to "be in a state of 'out of control' that they are incapable of knowing, unable to participate, unable to disagree, and ultimately difficult to resist." In this regard, it is necessary to construct a scientific algorithm governance mechanism, introduce an algorithm impact assessment and algorithm audit system, and eliminate adverse consequences such as algorithm discrimination and information cocooning. The fact that the results of the emergence, in part, are unpredictable also affects accountability. For general tools, the user can exercise effective control over them, so when the tools under his control cause damage to others, the user should be liable, and there is a clear causal relationship between the damage result and his control behavior, which is predictable. Humanoid robots are significantly different, and due to the emergence of human robots, the ability to control humanoid robots by humans is far less than that of ordinary tools. In the context of such a technology, where the likelihood of prediction is greatly reduced, it would be clearly inappropriate to pursue liability for damages in accordance with the traditional principle of fault liability. Penetrating the veil of technology, identifying the role and role of human subjects in the implementation of humanoid robot behavior, and properly distributing responsibilities in combination with the needs of industrial development will be the tasks that legislators need to complete in the future.

epilogue

After going through the three eras of steam engine, electrification and informatization, mankind is entering the era of Industry 4.0 - the era of intelligence, and the basic technology and application of the intelligent era are artificial intelligence and humanoid robots. Where humanoid robots are going in the era of intelligence is a test of human wisdom and determination. This article argues that humanity should uphold an anthropocentric stance: "We can insist on a culture that preserves, not destroys, a culture that complements, rather than replaces, humanity." We can achieve and afford a world ruled by people, not machines. "Embodiness enables humanoid robots to enter human domestic life, and the emergence of humanoid robots with powerful abilities that distinguish them from humans, and based on the position of anthropocentrism, we should position humanoid robots as objects in legal relations, rather than recklessly elevate them as subjects or restricted subjects. Standing on the moral high ground and treating all things equally will not give us a better perspective in dealing with the relationship between humans and humanoid robots, on the contrary, fraternity and pan-love without purposeful guidance will lead humans astray, humanoid robots are neither the same kind of human beings, nor are they legal subjects who interact with human beings on an equal footing, they are and can only be good tools for human beings to create a better world. No matter how technology develops, we should remain vigilant and rethink automation technology.

(This article is from the 3rd issue of Eastern Jurisprudence, 2024)

Thematic Coordinator: Qin Qiansong