Generative AI Policy
Number X, Inc. (the “Company”) uses generative AI in part of its work. Output from generative AI can contain errors or content that may infringe the rights of others. For this reason, the Company uses it on the premise that a person checks the results.
Scope and Basis
This policy applies both to the Company’s internal work and to work we undertake for clients. It is based on the AI Guidelines for Business (Version 1.2) issued by Japan’s Ministry of Internal Affairs and Communications and Ministry of Economy, Trade and Industry, and on the Act on the Protection of Personal Information and other laws.
Tools and Purposes
The Company mainly uses the following generative AI tools in its work.
- Claude (including Claude Code)
- ChatGPT, Codex
- Gemini
We use them to draft research, create draft text and images, write and review code, and help prepare documents. When using these tools, we follow each service’s terms of use and usage guidelines.
Handling of Client and Personal Information
As a rule, the Company does not enter personal information it has obtained, or clients’ confidential information, into generative AI.
If such input is unavoidable, we first confirm, through the service’s terms and settings, that the input will not be used to train the AI. Before entering personal data, we follow the procedures required by law, such as obtaining the individual’s prior consent.
When we build generative AI into a client’s system as part of an AI implementation project, the contract sets out how input data is handled, the scope of responsibility for output, and whether any work is subcontracted.
Review of Deliverables
Deliverables made with generative AI (articles, images, code and so on) are checked by a person before they are delivered or published. We check the following three points.
- Facts: whether figures, dates, proper nouns and sources are correct
- Copyright: whether the output resembles existing works, and whether materials may be used
- Quality: whether it fits its purpose and is free of typos and unnatural wording
Generative AI can produce incorrect information that sounds plausible. We do not include content in deliverables if its source cannot be confirmed.
Respect for Copyright and Other Rights
We check output so that it does not infringe others’ rights, such as copyrights, trademarks, portrait rights and personal information. We do not use content that may infringe such rights.
Transparency
If a client asks, we will tell them whether generative AI was used for a deliverable, and explain where it was used as far as we can.
Review of This Policy
We review this policy every six months. We also review it whenever generative AI technology, laws or guidelines change, and we announce any changes on this page.
Inquiry
Questions about this policy can be sent to the contact below. For how we handle personal information, please see our Privacy Policy.
Responsible person: Tsubasa Yasuda, Representative Director, Number X, Inc.
MAIL:info@number-x.jp
Established November 13, 2026