Deloitte Tohmatsu Group announced on September 29 the results of a survey on fraud risks conducted among 397 companies in Japan. While the adoption of generative AI is progressing, the findings reveal that the responsibility for managing risks is currently being left to individual users.
SecurityDeloitte Tohmatsu Group
74% of Japanese Companies Lack Measures or Visibility Regarding Personal Use of Generative AI, Deloitte Survey Finds
This article is a translation. Read the Japanese original
The survey showed that only 9% of companies responded that they are "not utilizing AI" for operational efficiency, indicating that the vast majority of companies have already integrated AI into their business operations.
Regarding AI risk mitigation, "formulation and dissemination of guidelines" reached 72%, and "regular training and awareness raising" reached 54%, showing a certain level of effort. On the other hand, companies that record and store input and output data were limited to 11%, and companies that periodically or manually audit output results were only 8%.
Deloitte stated that in response to risks such as leakage of confidential information and errors in output, companies are required to establish organizational risk management systems that can monitor usage status and enable early detection and response to problems.
Sources
- 7割の企業が生成AI「個人利用」への対策せず デロイト調査 （ITmedia AI+、2026-09-29）