This is mainly reflected in two levels. The first production level is machine production content, creative works, and patent applications. The second is the matching level, which is the automated push problem, which can push the content and advertisements that users want to see most. Since artificial intelligence as a technology has been continuously deepened, the production level has increased, and some competitive relationships have begun to appear. Many new legal issues have also emerged in this process. Laws are needed to confirm its legitimacy and to resolve conflicts. This kind of conflict happened once 20 years ago, which is the stage of illegal rise. The consequence of this process to this day is to confirm the legitimacy of the new economy. Now this trend is going very fast, and the law is trying to reconcile contradictions and promote this process.

Now we may be facing a second conflict. It may be the beginning. It is difficult to predict how long it will last, but at least we can try to summarize some of its characteristics. It is essentially no different from the first conflict, but still depends on the entrepreneur's desire for factors of production. The question now is, if machines join, can they produce some new patterns, open up new tracks, create a new form of media, and whether the form of mobile Internet that we are familiar with will also change, and so on. This will also lead to legal problems.

The emergence of artificial intelligence has brought some panic. Some economists are always talking about "machine substitution". Will human beings lose a large number of collective unemployment? But I think the core contradiction that the law is trying to solve is the contradiction between the platform of machine production and the platform of human production.

In other words, people may still have work to do, because there will always be platforms trying to mobilize such manpower, but the problem now is the emergence of a new platform, which claims to be a new species. This new species says that what we are producing now is completely produced by a fully automatic machine and does not rely on manpower at all. In this case, manpower now seems to have become a new cost, so the legal problem arises from the conflict between the production models of these two different platforms.

The demand for machine production is very consistent with the logic of the "illegal rise" of the early Internet. Machines are first and foremost dependent on human activity, and they cannot be created out of thin air. In addition, artificial intelligence companies, including face recognition, need to have a large number of face data sets for training. This data set is crawled from the Internet, and there are a large number of face photos on the Internet. This machine learning itself relies on the labor process provided by the human. This process will lead to a problem, that is, the platform economy will reduce the cost of information production and matching, there is a competitive advantage. But for producers, everyone is using the Internet, and labor costs are coming up again; everyone is starting to produce spam, and users may get very little incremental value from these spam.

If many people have to invest a lot of time to start rolling, it will be wasteful, because the things produced will not necessarily be pushed to those who need it. So in this sense, machine production is actually an impact on the original expert system. Experts must not only produce high-quality information, but must also compete with spam.

Start-ups are all based on this logic. The development of new track virtually provides a lot of pressure on the original platform or producer. The whole environment has gradually become a process of continuous interaction with the subject. Data is no longer purely as consumer data, but also includes our life and behavior data. The offline world or smart city is a "meta-universe". Moreover, users will continue to interact with non-human elements, what you use, what you buy, and what you see. In this interactive process, the machine will continue to learn and make a comprehensive simulation.

What are the legal issues facing this process?

The first is the problem of intelligent aggregation infringement. Automatic video aggregation and face changing may be illegal. There is also the problem of fake news. For example, there are many water troops on the Internet, or the account itself is controlled by some robots. It can mass produce real news, fake news, and even some accounts are controlled by machines.

This involves different levels of issues. The first is the legality of machine production. For example, whether news must be produced by people or machines, including justice, is the same. In form, a human judge is required to make decisions, or machines can be allowed to participate in decision-making. These are similar issues.

If we admit that machine production is no problem, we need machine production. Then the next problem is the competition between machines and humans. Should humans be attached to the machine production process, or another track?

On the surface, the legal issue seems to be a neutral issue, but it is not entirely. Formal legal issues sometimes cover up and reflect production relations. The law seems to operate in a vacuum. We only need to read the law or read the case, but only when it is connected with the social production process, we can Understand how the legal problem came out and who wants to use the law to achieve what purpose.

So what impact will machine production have on consumers? First of all, whether machine production can increase our wealth? This problem is prominently reflected in some virtual equipment, virtual property, that is, some of the things we have in the virtual world. Platforms or game companies will say that this game equipment cannot be given to you. In other words, you can only own these things in the meaning of the game, but you cannot permanently own them without this scene. So in this sense, it is only part of the service and cannot add additional wealth to you.

In addition to having an impact on consumers, it also has an impact on the production structure of the entire information market. If machines become competitors, human producers will have to develop more types of services. At this stage, it seems that some jobs have been replaced by machines, but people will develop more new types of jobs. But the crux of the problem is how to control the competitiveness of the market structure. If the purpose is to rely on machines to reduce human production costs, it may bring some countereffects. Secondly, from the perspective of production, we will initially try to follow the past Internet model, such as free first and then charge, in order to maintain scarcity. If this market can operate effectively, many elements must be in a state of scarcity, otherwise there is no way to make a profit, so in the process of a large number of machine production, how to improve the quality is always a problem.

There is also the issue of automatic content push. Speech is not a consumer market, it is an expressive market, and expression itself is meaningful. A creative society needs to make people speak. If it is over-commercialized or pushed through machines, it will narrow the information. If the algorithm push is inevitable, then you can only take the initiative to make more efforts to improve the speech environment.

There is another problem with machine control of the market. Machines can create a lot of content, but this content ultimately needs people to consume, not machines.

Therefore, we still regard people as consumers or workers, so in the process of human-computer mixing, behavioral data may be more valuable. Two types of data are distinguished functionally, the first type of function is display data, and the other type is auxiliary data, which can help the market flow more orderly and safely, and help these display data better match.

In the past, we talked about many issues, including data property rights, data ownership issues, and transaction issues. We talked about the first type of data, such as questions and answers on Zhihu, or videos on trembles, and information on microblogs. This type of data may have been generated by humans, but if the platform now starts to use machines to produce these data, what will happen? These things may slowly cease to be competitive assets and their importance as a legal issue may diminish. But the second type of data becomes more important, which is reflected in the competitive advantage of large platforms. There are many users and many producers on the big platform. Now machine production is introduced, but human producers will not quit, and they may develop more services, which can also be transformed into consumers to consume more machine information.