AI explains. JMA decides.
SHINDO uses AI to put official information into plain words. It never decides whether you are warned, and it never sees your personal data.
SHINDO uses AI to put official information into plain words. It never decides whether you are warned, and it never sees your personal data.
| Feature | What the AI does | Input |
|---|---|---|
| Earthquake and tsunami explanations | Writes a short summary in English, Japanese and Spanish after JMA's report | JMA epicentre, depth, magnitude, observed intensities, tsunami status |
| Weather and typhoon briefings | Summarises current warnings, forecasts and storm tracks | JMA warnings and typhoon parameters, Open-Meteo forecasts |
| Transit and news translation | Translates operator notices and headlines | Published notice and headline text |
Model provider: Anthropic, PBC (Claude models), called from our server in Tokyo.
Every AI-written text in the app carries an "AI" label. When the AI service is unavailable, the app shows a plainly worded summary built from the official data and labels it "Summary", not "AI". The official numbers — magnitude, depth, intensity, warning level — are always shown next to the text and take precedence.
Language models can make mistakes, including wrong place names or misread figures. We constrain them to the official data provided, check output format automatically, and correct reported errors. Do not rely on AI text alone for safety decisions; follow JMA and local authorities.
If an AI explanation was wrong or misleading, tell us at support@shindoalert.com with the time and the event. We review every report.
The transparency obligations of Article 50 of the EU Artificial Intelligence Act (Regulation (EU) 2024/1689) for AI-generated text on matters of public interest; Japan's Act on the Promotion of Research, Development and Utilization of AI-Related Technologies (2025) and the AI Guidelines for Business (AI事業者ガイドライン); Apple's App Review Guidelines on disclosing data shared with third-party AI; and the privacy laws listed in our Privacy policy. SHINDO's AI features are not used for emergency dispatch or to evaluate people and are not a high-risk AI system under the EU AI Act.