Reporting AI responses

AI can make mistakes — language errors, mistaken advice, or wrong facts in the examples. Please use Geeks in Kyoto as an instructor and a practice resource, not as an authority. It's for language learning; don't rely on it for important matters such as legal documents, medical situations, or business communication. For those, consult a qualified professional or a native speaker with expertise in your domain.

What you can report

If you see an AI response that is particularly concerning, you can report it for human review. That includes, but isn't limited to:

Technical issue / Bug — garbled text or speech, broken delivery of prepared or interactive content, or anything else that simply doesn't work.
Language or explanation — vocabulary or grammar mistakes, wrong advice on usage, confidently stated wrong facts, made-up cultural information, or invented references.
Offensive or inappropriate — sexually suggestive, rude, disrespectful, biased, or otherwise uncomfortable content.
Legal, copyright, or harmful — copyrighted material; impersonation of real people or fictional characters; references to third-party brands, trademarks, or institutions unaffiliated with Geeks in Cities Ltd (including, but not limited to, JLPT, NHK, and others); responses that appear to include passwords or other secret data, or attempts to use the AI tutor to communicate with other users; threats, doxxing, or any material affecting the safety of children.

How your report is handled

Reports are dealt with according to severity. Bugs and language or explanation errors are considered for change as part of ordinary product improvement. Offensive content is reviewed promptly. Legal, copyright, or harmful reports — although extremely unlikely — are reviewed by a person, initially at Geeks in Cities Ltd and potentially escalated to our AI provider and other relevant parties.

We reserve the right to withdraw the service from any user who, following review, is found to be manipulating the AI — through prompting or other means — to make it produce responses that fall within the categories above.

What we keep when you report. By reporting an AI response for human review, you agree that a record containing the AI response, the conversation leading up to it, your account details, and your reported reason will be kept by Geeks in Cities Ltd. We retain these records only for as long as necessary to action your report and to spot recurring issues — typically up to 24 months. You can request deletion of your account data at any time via [email protected], though we may keep anonymised pattern data afterwards to improve safety.