A survey by アクシスコンサルティング of 647 employed people nationwide found that concerns about AI use centered on practical risks such as accuracy and information leaks, while concern that part of their work would be replaced was relatively low.
アクシスコンサルティング conducted the “Survey on Challenges and New Tasks and Roles in AI Use” among employed people nationwide in their 20s to 60s, focusing its analysis primarily on a total of 647 people who use AI at least once a month. The survey was conducted from 11 to 12 September 2026, and 603 respondents were included in the main analysis. A key characteristic of the results was that concerns centered on practical quality and risk.
The leading concerns and challenges in AI use were “accuracy” at 49.4%, “information leaks” at 41.8%, “declining decision-making ability” at 31.2%, and “copyright and related rights” at 30.7%. The least common concern was that “part of one’s own work would be replaced,” at 5.8%. The results highlight concerns directly linked to practical quality and risk.
The most common newly created task or role was “reviewing output and quality control” at 40.8%, followed by “information management and rule development” at 30.5% and “instructing AI and designing its use” at 27.2%. The introduction of AI use appears to be creating new management and design-related tasks, such as output quality control and rule development.
There was a relationship between sources of concern and the emergence of new tasks. Among those concerned about accuracy, the proportion reporting the emergence of “instructing AI and designing its use” was high at 41.9%, compared with 12.8% among those with no such concern. Among those concerned about information leaks, “reviewing output and quality control” was high at 58.7%, while among those concerned about copyright, “information management and rule development” reached 48.1%.