Historically, the iteration of "hacker" to "artificial intelligence hacker" was accompanied by the rapid development of science and technology, such as computer, internet, big data and artificial intelligence.Nowadays, "artificial intelligence hacker" is a human-computer interaction, which is neither a person nor a thing. It can imitate human beings, interfere with human cognition, and intelligently invade and destroy network system vulnerabilities for the purpose of designers or decision makers.The main feature of "artificial intelligence hacker" is that it can rely on intelligent algorithms to learn autonomously, find network system code vulnerabilities and strengthen distributed attacks.Some scholars divide artificial intelligence technology into three stages:weak artificial intelligence, strong artificial intelligence and super artificial intelligence. Some scholars even suggest that strong artificial intelligence be given the legal subject status ethically, because strong artificial intelligence algorithm has independent "machine meaning" expression ability and has emotional connection with human beings.Obviously, this way of empowerment violates the principle of "humanism". Also, the current legal subject includes natural person, legal person, unincorporated organization, "artificial intelligence hacker" does not belong to any kind of subject, abruptly equating the rational expression of legal subjects with the "machine meaning" of artificial intelligence algorithm instructions can easily lead to difficulties in algorithmic justice legal evaluation and civil legal behavior construction of "artificial intelligence hacker" behavior, interfering with our judgment on the nature of "artificial intelligence hacker". The legal attribute of "artificial intelligence hacker" should be judged based on legal rights and obligations. In essence, "artificial intelligence hacker" is a natural person who uses network media to commit network infringement or crime through artificial intelligence algorithm technology.The core of "artificial intelligence hacker" is a set of mechanism for making decisions through computer code setting, big data operation and automatic judgment of machines. "Artificial intelligence hacker" is not a proper legal subject in terms of responsibility, but has a special legal attribute of "personality tool".The intelligent attack of "artificial intelligence hacker" is manifested by the automatic execution of the algorithm program, but the design and operation of the program belong to people in real economic life, which also fully conforms to the adjustment scope of indirect infringement in law. For criminal acts or infringement of "artificial intelligence hackers", the hidden subject and object can be regulated by law should be found through piercing the "veil" of "artificial intelligence hacker", and the "penetration" method should be used for the effective regulation of illegal behaviors of "artificial intelligence hackers" from the dimensions of ethics, technology and law.