The number of cases of fraudulent accounting of companies has been increasing year by year, which has become a major social problem. Until now, fraudulent accounting has been detected by relying on the experience and intuition of practitioners, however machine learning methods have been attracting attention in recent years. This research group is also proceeding with the development of a model with higher accuracy than the methods of previous research and those that have been put into practical use by an audit firm.

In this project, we will build and socially implement a novel fraudulent accounting detection method using high-precision machine learning (AI) that can be deployed not only to some practitioners such as certified public accountants but also to a wide range of interested stakeholders including general investors, toward, exploring a new form of co-creation with advanced technologies.

For that purpose, in addition to constructing a highly versatile fraud accounting detection model and considering the needs and requests of the practitioners for practical use, we will clarify the legal issues when used by company directors and audit corporations, showing the theory of conversion, interpretation and legislation, and the new legal framework for solving it will also be clarified.

The method and software to be developed through this project are planned to be released free of charge, which not only contributes to strengthening corporate governance and significantly reducing auditing costs, but also alleviates information asymmetry among stakeholders. Accordingly, we aim to form a sound capital market through new co-creation of advanced technology and society.