Takako Hashimoto, Director of Chiba University of Commerce Research Center for Economics

— Will the job of "accounting auditor" disappear?—

In December 2015, Nomura Research Institute, Ltd. (hereafter, Nomura Research Institute) announced a joint study with Michael Osborne Associate Professor professor of artificial intelligence at Oxford University, that "in 10 to 20 years, 49% of the Japanese workforce will be replaceable by artificial intelligence and robots, etc." [1]. Nomura Research Institute also introduced "100 occupations that are likely to be replaced by artificial intelligence and robots, etc." One of these is "accounting auditor." In fact, relatively simple accounting tasks such as issuing invoices, bookkeeping, and preparing financial statements are already being realized by systems such as ERP (Enterprise Resources Planning). Automating routine work is what computers are best at, and many companies are reducing their workforce. On the other hand, accounting work is made up of management accounting such as sales management, cost management, and profit management, and financial accounting, which publishes the company's management performance to the public, and requires various knowledge and evaluations. It has generally been thought that tasks that require knowledge and experience cannot be easily replaced by information technology. However, recent advances in artificial intelligence technology have made it possible to replicate the specialized knowledge and experience required for accounting work on a computer. This has led to hopes of reducing the number of people working in accounting. For example, Free Inc., which develops the cloud accounting software "freee," estimates that a company with 300 employees could get by with just 0.8 accounting staff. Many accountants and tax accountants feel a sense of crisis about this situation.

—What kind of accounting work can be done with artificial intelligence?—

Let's think about what kind of accounting work can be replaced by artificial intelligence. By using cloud accounting software such as "freee", it is possible to automatically create financial statements and tax return documents, and to grasp the business situation in real time. In addition, it is expected that artificial intelligence will be used for "automatic guessing of account items" and "detection of fraudulent accounting". Cloud accounting software automatically obtains a company's transaction details. The data obtained is "date", "transaction amount", and "transaction details such as the counterparty of deposits and withdrawals". When a company creates an accounting book, it sets and registers accounting account items based on these three pieces of data, but by using artificial intelligence technology, it is possible to guess these account items. Even if it is a special proper noun (an uncommon name such as "XX Co., Ltd." or "XX Office"), if you manually register it in the book several times, the machine learning technology of artificial intelligence will automatically guess it. Learning will be even easier and more accurate by utilizing Google's gmail exchanges, search history, calendar, etc. It is expected that "automatic guessing of account items" will significantly reduce the bookkeeping work such as bookkeeping performed by tax accountants and accountants.
Regarding "detection of fraudulent accounting," in November 2017, Ernst & Young ShinNihon, a major auditing firm, announced that it had put into practical use an algorithm to detect anomalies in accounting entries using artificial intelligence [2]. This algorithm uses artificial intelligence to learn transaction patterns from accounting entry data and automatically identify abnormal entries. Traditionally, auditors set search conditions based on certain hypotheses to extract abnormal entries that fit these conditions. On the other hand, this method uses machine learning to learn tens of thousands of patterns of accounting entries, and the AI automatically detects abnormal entries such as overstating revenues and understating expenses. Pattern recognition is also a task that computers are good at, making it possible to significantly reduce the amount of manual checking work that was previously required. Accounting work that required specialized knowledge and experience is also being replaced by artificial intelligence. With the spread of artificial intelligence, it is expected that the amount of specialized knowledge required by tax accountants and accountants will decrease, and the work and skills required for accounting work will change significantly.

—To survive in this era of change—

The rise of artificial intelligence technology is inevitable. We need to grasp these changes and respond appropriately. Rather than feeling a sense of crisis unnecessarily, it is important to think about what we should learn and what added value we should create in order to survive this era of change. It is important to remember that artificial intelligence does not evaluate or judge, but merely learns from a huge amount of data and classifies it based on that. When an exceptional event occurs, artificial intelligence may not be able to make an accurate judgment. It is also difficult to make flexible judgments according to the situation or provide advice or consulting that resonates with people. Now that we are entering an era where artificial intelligence can reproduce experience and knowledge, it is necessary to properly acquire basic accounting knowledge and develop the skills to make flexible judgments, decisions, and advice. The author believes that this will be an important added value for the personnel who will be responsible for accounting work in the future.

References
[1] Nomura Research Institute: 49% of Japan’s workforce could be replaced by artificial intelligence, robots, etc.
https://www.nri.com/jp/news/2015/151202_1.aspx
[2] EY Japan BMC puts AI-based accounting anomaly detection algorithm into practical use
https://www.shinnihon.or.jp/about-us/news-releases/2017/2017-11-06.html