Primary survey · Source language: Japanese
The burden of checking AI differs by how it is used
19.5% of generative AI users in Japan answered that checking information or choosing answers is a burden. The share is higher among those who check frequently.
- Who:
- Generative AI users aged 15–69 across Japan
- Scope:
- Generative AI users: 1,995 in 2025 and 3,669 in 2026 / Japan
- Source published:
- 2026-09-24
- Period observed:
- February 2025 and February 2026
- Editorial check:
- 2026-10-23
Only the source ID is saved in this browser. Nothing syncs across devices.
What was confirmed
- Dissatisfaction with misinformation rose from 34.8% in 2025 to 40.4% in 2026.
- The burden of checking and choosing was 19.5% of all users, and 28.2% of those who always fact-check.
Who, and in what situation
Compared surveys of February 2025 and February 2026, and also analyzed the 2026 checking burden.
What they needed
Judging correct information and choosing answers that fit the purpose. Whether it is a burden varies by user.
What they tried
Some users find it useful while checking the content. The actual working time of the burden was not measured.
What this source lets you read
Rather than lumping it together as "AI fatigue", it is worth looking closely at the work of people who use it for important purposes.
What to check next
Support that lowers the checking burden would help people with which uses and frequencies?
Limits of this source
- Background material showing the state in February 2026.
- The relationship between checking frequency and burden is a correlation. Demand for paid services was not measured.
Read the primary source
Links use the original title or a HAKOMIRO source description. Japanese originals retain their published wording.
生成AIの不満「誤情報」40%に増加、確かめながら使う姿勢の広がりも (in Japanese, opens in new tab) ↗
NTTドコモ モバイル社会研究所 /
An online registered panel allocated to match population composition. The whole survey had 7,527 and 7,223 respondents.