'Sophia' replied: 'You don't have to worry about our robots, how do you humans know that you are human?' Last year, the artificial intelligence robot 'Sophia' became the world’s first robot to be given legal citizenship.
At present, great advances in science and technology are driving the rapid development of artificial intelligence. What are the challenges to the legal system brought about by the profound changes in production and lifestyle brought about by artificial intelligence? How should the current legal system be adjusted and responded to?
Does artificial intelligence produce intellectual property?
In the light of the twilight, I know her lovely soil, which made my heart a captive...' The creator of this verse is not the 'human' in the traditional sense, but the artificial intelligence product 'Microsoft Little Ice'. In May 2017, the “Microsoft Little Ice” poetry collection “Sunshine Lost Glass” was published. As the first poetry collection entirely created by artificial intelligence in history, its publication brought a new problem—artificial intelligence products. Is there intellectual property?
According to Cao Xinming, director of the Intellectual Property Research Center of Zhongnan University of Economics and Law, according to current laws, intellectual property achievements refer to 'human achievements' and artificial intelligence cannot be the subject of rights in the sense of intellectual property rights. 'But if the 'artificial intelligence 'Creation activities' are analogous to the 'computers' of scientific research, that is, artificial intelligence products are regarded as intellectual achievements created through artificial intelligence, and artificial intelligence products do indeed possess certain attributes of 'intellectual property works'.
The key to the problem lies in the legal characterization of 'artificial intelligence'. ' Cao Xinming stated that at present the academic community has two views on this issue: 'tools' and 'virtual man'. 'Tools' is to treat artificial intelligence as human's creation. And rights object; 'virtual person' is the attribute that the law sets part of 'person' to artificial intelligence, giving it legal qualifications that can enjoy some rights.
'Even if it is acknowledged that artificial intelligence products have intellectual property rights, the ownership of their rights is a question that needs to be answered.' Cao Xinming believes that if artificial intelligence is regarded as a 'tool', the rights of artificial intelligence products can be attributed to the design developer, or ownership. People, or rights holders, and multiple rights holders. If artificial intelligence is viewed as a "virtual person," artificial intelligence artifacts can be viewed as a "suffocation" in the sense of civil law. For example, artificial intelligence is considered as a "mother." Chicken', then the artificial intelligence product is the 'egg' under the 'hen', and the 'egg' is owned by the owner of the 'hen'.
In addition, creating artificial intelligence products often involves 'deep learning' through a number of programs, which may collect and store a large amount of intellectual property rights that others have already enjoyed. This may constitute an infringement of the intellectual property rights of others. Cao Xinming believes that In the case of such alleged infringement of intellectual property rights, who should bear the responsibility should be a new issue.'
Can artificial intelligence replace the judiciary?
In recent years, the application of artificial intelligence has gradually deepened in the judicial field: In December 2016, the Beijing Court Intelligence Judgment System named “Judge Rui” went online to provide judges with accurate information on handling case specifications and sentencing analysis, and to use big data to advance the application of laws. In May 2017, the country’s first “Criminal Case Intelligence Aided Casework System” was born in Shanghai. After “deep learning” of tens of thousands of criminal case files and document data in Shanghai, it already had preliminary Evidence information crawling, checksum logic analysis capabilities...
'Using artificial intelligence can help the judiciary get all the precedents of similar cases and referencing rules such as laws, regulations, and judicial interpretations, so as to reduce their workload and promote the accurate application of the law.' Researcher at the Institute of Law of the Chinese Academy of Social Sciences Zhiyong Zhu It is believed that through data collection, collation, analysis, and synthesis, artificial intelligence has a lot to do in promoting the judiciary in accordance with the law, comprehensively, standardizing the collection and review of evidence, harmonizing judicial standards, and assisting justice.
However, does this mean that artificial intelligence will replace the judiciary and achieve an independent case? Obviously not.
'Artificial intelligence is just an aid to judicial justice. It must not be placed in the cart before the horse. This is a basic principle that we should always remember.' According to Professor Ji Weidong, Dean of the Kaiyuan Law School at Shanghai Jiaotong University, if we artificially rely on artificial intelligence to automatically generate judgments Depending on the deviations of big data correction legal decisions, it is inevitable that a multiple structure of the trial subject will be formed. In fact, the subjects and the judiciary, such as programmers, software engineers, data processors, etc., will jointly implement the judicial situation.
In addition, if artificial intelligence is used beyond the scope of ancillary measures and is fully applied in trial cases, then it is possible to mislead justice. 'Ji Weidong believes that in the case of factual twists and turns, complex interpersonal relationships, and ethical and emotional factors, How to make judgments based on jurisprudence, common sense and human feelings, and make proper rulings is actually a subtle art that needs to rely on the judge's rational comprehensive analysis. 'Even if artificial intelligence embeds probabilistic procedures and has deep learning capabilities, it is difficult to guarantee that Fair and reasonable, convincing case referee. '
Zhizhen Feng also believes that, judging from the current development situation, artificial intelligence has not yet replaced the possibility of the judiciary, especially as a legal action involving emotions and rationality, norms and values. If given to artificial intelligence, this is legally and ethically important. , it is difficult to get support. 'should prevent the formation of 'path dependence' of artificial intelligence, the more developed artificial intelligence, the more emphasis should be placed on the professional ethics of the judiciary.' Zhi Zhenfeng said.
How to determine the liability of artificial intelligence infringement?
In November 2016, at the 18th China Hi-Tech Fair held in Shenzhen, a robot named 'Little Chubby' suddenly malfunctioned and broke some booths without instructions. And it caused one person to get hurt.
The increasing popularity of artificial intelligence applications, the issue of infringement liability and commitment caused by it, is a new challenge to the current infringement legal system.
From the current legal point of view, the main body of tort liability can only be a civil subject, artificial intelligence itself is still difficult to become a new subject of tort liability. Even so, the determination of artificial intelligence tort liability also faces many practical problems. ' Professor at the School of Law, Tsinghua University According to Cheng Xiao, after the infringement, who should be responsible for the artificial intelligence, it seems that there is no legal dispute. 'However, the specific behavior of artificial intelligence is under the control of the program, and in the event of infringement, it is The owner or the software developer is responsible for it. It is worth discussing.
Similarly, when a driverless car causes harm to others, it is the driver, the owner of the motor vehicle, or the car manufacturer, the developer of the automatic driving technology, who is responsible for it? Cars develop special tort liability rules? These issues are worth further study.
'In reality, the principle of attribution of liability for infringement of artificial intelligence may be more involved in dangerous or no-fault liability.' Cheng Xiao believes that, for example, the harm caused by driverless cars, whether from product liability or the liability of motor vehicle traffic accidents, No fault liability can be applied. But the future need to consider is whether the use of artificial intelligence technology itself is a highly dangerous operation (such as drones), thus determining whether to apply high-risk operations responsibility.
'Currently, the judgment of causality, faults and other factors in the tort liability of artificial intelligence has also become increasingly complex.' Cheng Xiao also cited examples of previous exposures of 'applicability of big data to killing' and 'algorithm discrimination' due to the code's Opaqueness, combined with the self-learning and adaptability of the algorithm itself, makes it very difficult to “put the algorithmic blame on the developer”.
According to Cheng Xiao, the new problems and new challenges brought about by artificial intelligence will be taken for granted in the study of the legal system and will take the initiative for future judicial practice. 'Artificial intelligence has arrived, but it is not distributed in every area of production and life. We should not wait until the future is evenly distributed. When artificial intelligence has been fully integrated into all aspects of production and life, we only think of it from the law. " Cheng Xiao said.