User-Posted Content Safety and Integrity Risks
First written in FY2025. 1 of 2 entries say "Mitigated", first in FY2026. Latest entry: FY2026.
The whole theme swung hardest in 2025: 4 of 21 companies on the "Worse" side.
1 other company's latest entry also says "Mitigated".
- FY2026MitigatedAnnual report
「One hour after a user post is submitted, we use AI to check every post for public order and morality and hygiene-related issues. In addition, to verify AI accuracy, at the beginning of each month we randomly sample articles published in the preceding month and conduct a subsequent review by the responsible personnel.」
In addition to conventional post monitoring and visual checks, the company introduced comprehensive AI checks and monthly sampling reviews. - FY2025First statedAnnual report
「However, although we have strengthened our monitoring framework, if inappropriate content is posted on the service and we fail to detect it, if our response is delayed, or if a third party raises concerns about cooking methods or the handling of ingredients, our credibility may be damaged」
The company cites a post-monitoring system, manual visual checks, checks of initial posts, sample checks, and checklists for allergy and cooking-related risks as countermeasures.