• English
  • Japanese

Guideline for Generative AI Usage at Keio University

1. Introduction

1.1. Purpose of This Guide

Generative AI has the potential to dramatically improve the efficiency of learning, research, and work. The purpose of this guide is to correctly understand the risks of using generative AI, ensure information security, and maximize its benefits. Each member of the Keio University is required to comply with the content and use it responsibly.

For the use in classes, reports, and assignments, please be sure to check "Announcement on Using ChatGPT and Other AI-Generated Content (May 15, 2023)" in conjunction with this guide, and use it appropriately under the policy indicated by the course instructor.

1.2. Target Audience

All students, graduate students, and faculty and staff members (including those at integrated education schools and university hospitals, regardless of full-time or part-time status) enrolled at Keio University.

Currently, for Google Workspace for Education, age restrictions etc. are applied based on the Terms of Service and the tenant management rules under the Keio University contract.

2. Precautions and Prohibited Matters for Use (Common to All Users)

When using various AI services (including those both inside and outside corporate contracts), the following matters must be strictly observed.

2.1. Handling of Information

Exercise sufficient caution when handling personal information, privacy-related information, confidential information, etc. Even when using corporate services contracted by Keio University (Chapter 3 "C" described later), careful handling is required, and especially for services outside corporate contracts, do not input personal or sensitive information through prompts or file attachments.

2.2. Basic Principles for Appropriate Use

  • Thorough Fact-Checking: Generative AI may produce information that is not based on facts (hallucinations). It is strongly recommended to always verify generated information against primary sources (original papers, reliable materials, etc.) and perform fact-checking.
  • Ensuring Originality and Creativity: Information provided by generative AI should be considered as reference material only, and the final output must be constructed based on the user's own thinking. It is important to add one's own perspective to the generated information to ensure originality and creativity. For the use of generative AI in class assignments, please be sure to check "Regarding the Use of ChatGPT and Other Generative AI at Keio University (May 15, 2023)" and use it appropriately according to the course instructor's policy.
  • Ensuring Transparency and Reproducibility in Research: While generative AI makes it easy to edit and process images and data, and to summarize information, using generated information as is may inadvertently lead to research misconduct, so sufficient caution is required. Please fully understand that researchers themselves are responsible for ensuring the transparency and reproducibility of their research and for their research outcomes.
  • Compliance with Rules for Each Research Activity: When including generated AI outputs in research, it is required to follow the policies of the academic journal or society to which it is submitted, such as clearly indicating the AI tool and model used, prompts, and the source of the generated content. Additionally, in each research activity, all related researchers should strive to share a common generative AI usage policy and ensure a unified response.
  • Consideration for Copyright: When handling copyrighted papers or articles, etc., be careful not to deviate from the scope permitted by copyright law (e.g., use for learning/research purposes). Depending on the AI service used, there is a risk that copyrighted works handled may be used for purposes other than intended against your will.
  • Confirmation of Generated Output: There is a possibility that expressions closely resembling existing copyrighted works may be generated unintentionally, and using generated outputs similar to existing copyrighted works without permission from the copyright holder constitutes copyright infringement. Please make an effort to confirm whether there are similar expressions before using the generated information.
  • Ethical Use: Do not use for purposes that promote discrimination or prejudice, defame others, or violate public order and morals.

3. [Important] Classification of Various AI Services and Prerequisites for Use

Generative AI services are diverse, but the information security risks differ between "those provided under Keio University's management" and "those used individually." To protect input information (personal information, research data, confidential information, etc.), please be sure to understand this difference.

AI Service Classification [A] License-Free AI Services
(Services available without issuing an account or explicit contract)
[B] Individual/Personal Contract AI Services
(Services requiring account creation or license agreement)
[C] AI Services Based on Corporate Contracts
(Specifically, those contracted by Keio University)
Example Usage Forms The group of AI services such as ChatGPT, Gemini, Claude, etc., that can be used license-free.
Example: No restrictions on service use, or available simply by agreeing to the terms of service.
The group of AI services such as ChatGPT, Gemini, Claude, etc., that require account registration or an explicit contract.
Example: Registration of a user account, individual billing via credit card, etc.
(As of July 2025, in trial implementation)
Gemini, NotebookLM used under Keio University's contract. *Provided as the standard function version.
A comprehensive contract at the corporate level, available under Keio.jp accounts.
Risk of External Data Leakage or Use for Model Training There is a high risk of data leakage with user data, and concerns that it may be used for model training in the AI service. Opt-out (restriction of external leakage of user data) is standard, or cases where it can be selectively configured are mainstream.
*Careful confirmation is required as terms of use and settings differ for each service.
By default, external data leakage and use for model training are restricted and managed (premised on use under overall opt-out control).
Use in Administrative and Corporate Operations Refrain from use due to the risk of leaking sensitive information. Refrain from use due to the risk of leaking sensitive information. Available through functions within the corporately contracted service.
Use in Education (Classes) and Research Refrain from use for education (classes) and research. Use only after thoroughly confirming that there are no risks to individuals or university activities, such as the application of opt-out, in the terms of use for each AI service. Available through functions within the corporately contracted service.
Handling of Personal and Sensitive Information Avoid inputting (via prompt input or file attachment) personal information or sensitive information related to organizational operations. As a general rule, avoid inputting (via prompt input or file attachment) personal information or sensitive information related to organizational operations. Sufficient caution is required when handling personal information, privacy-related information, and confidential information. It is desirable to basically avoid unnecessary input (via prompt input or file attachment).

Regarding Google's "Gemini Free Upgrade Offer for University Students"
This is to upgrade the functions of an individual Google account by entering a Keio University email address (@keio.jp) as registration information. Unlike the corporate contract service managed by Keio University (C above), this is classified as B. an individually contracted AI service. Please carefully review the terms of use and pay attention to the precautions based on B above. (An announcement has been made that the 15-month student benefit ended on June 30, 2025. For the latest information and details, please refer to Google's page.)

4. Main Features of Gemini/NotebookLM Provided by Keio University (Trial Introduction as of July 2025)

As of July 2025, Gemini and NotebookLM provided under Keio University's contract within Google Workspace for Education have been trialed with standard feature versions, and by accessing from @keio.jp, it is possible to use AI services based on a corporate contract. While some paid features, such as image generation and more advanced models (Gemini Advanced), are not available, they can be safely used for learning, research, and work.

For detailed operating procedures, please refer to "About the Use of Gemini and NotebookLM."

4.1. Gemini

Supports a wide range of intellectual production activities such as text creation, summarization, translation, idea generation, and information retrieval assistance in a conversational format.

4.2. NotebookLM

A tool specialized in answering questions and summarizing based on the content of materials uploaded by the user (PDF, text files, Google Docs, etc.).

5. Responsibility

  • Generative AI is a "tool" to assist learning, research, and work.
  • The final judgment and responsibility for how AI-generated content is used and presented as an outcome rests solely with the user.
  • Keio University is not responsible for any problems (information leakage, copyright infringement, research misconduct, etc.) arising from the use of AI. Users must strongly recognize that they themselves will be held responsible.

Last-Modified: July 2, 2025

The content ends at this position.