The quantitative and qualitative differences between cybercrimes and traditional crimes have presented new challenges to the effectiveness of the traditional criminal justice system. Artificial intelligence technology, with its ability to analyze data, learn from patterns, and make automated decisions, is recognized as a powerful tool for legal applications and a dangerous means for criminal activity. The prevalence of cybercrimes committed by human agents and the emergence of these systems at two levels, advanced intelligent and semi-intelligent, have made the basis of attribution and responsibility in cybercrimes challenging. In computer science, intelligent systems are considered to have the power to analyze and autonomously execute a command, and are assumed to possess perception and autonomous action. Therefore, identifying the agent requires analyzing the causal relationship and questioning the criteria for attribution and identifying the position of the parties. Research findings indicate that in Iranian judicial practice, there are no cases directly related to AI responsibility, and in cases related to cybercrimes, courts typically seek to identify the person or persons who used technology to commit the crime. For example, in internet fraud cases, responsibility is attributed to the individual who used software or automated systems to commit the crime. Therefore, user responsibility is still based on fault, and manufacturer responsibility is absolute, and AI is not considered more than an object. The aim of this descriptive-analytical article is for the courts, by carefully considering the role of autonomy in determining responsibility and attribution, to determine the advanced intelligent AI agent in cybercrime, given its autonomous mode of operation, using a customary attribution model inspired by the theory of object responsibility, and even with the unique characteristics of advanced AI, it is appropriate to define an independent personality for it.