1. 基本理念
- 無断生成・盗用を許さない: 著作権および使用権を遵守し、他者の権利を侵害しません。
- 情報収集に生成AIを使用しない: 公的機関や信頼できる情報源を用い、生成AIは補助的なツールに限定します。
- 生成AIの適正活用: EGICの活動を補助する目的に限定し、生成物の管理とリスクの認識を徹底します。
生成AI利用に関するガイドライン
信頼できる情報共有と、AIとの、正しい、共存のために。
コレクトのナビゲーション
EGICは、生成AIの、存在を、否定しません。しかし、それを、悪用する行為は、許しません。これより、生成AIを、「正しく」利活用するための、当団体の、基本方針を、開示します。
2. 適用範囲と運用方針
本ガイドラインは、EGIC全職員に適用されます。盗用禁止を徹底し、AI生成を主体とするコンテンツは明確に示し、透明性のある運用を行います。
3. ファクトチェックの実施と管理
公式情報源を基にした2回以上のファクトチェック、2人以上の人員による検証、そして、必ず、人の目による最終確認を行い、情報の正確性を担保します。
4. 団体内での生成AI使用対策
団体内の生成AI使用は、翻訳や校正など補助業務に限定し、意思決定や情報共有には使用しません。使用目的と生成物の内容は、3日以内に、上層部へ、報告することを、義務付けます。
5. 違反行為への対応
違反が疑われた場合、速やかに調査・是正措置を行います。また、外部からの通報を受け付ける窓口（support@ewrs.jp）を設置し、適切な対応を行います。
6. 生成AIの利活用推進と信頼構築
生成AIの適正利用についての知識を普及し、倫理的な利用を推進します。情報収集に生成AIを用いることはありません。
7. 改訂と更新
本ガイドラインは、技術進展やリスク変化に応じて定期的に見直しを行います。
1. Basic Principles
- No Unauthorized Generation and Plagiarism
EGIC will not tolerate unauthorized generation or plagiarism of generated AI. EGIC will respect copyright and usage rights and ensure that the rights of others are not infringed. - Do not use generative AI to collect information
We will use public institutions and reliable information sources, and limit the use of AI generation to a complementary tool. We will never use generative AI as a means of gathering information for disaster prevention. - Appropriate use of generated AI
The use of generated AI will be limited to the purpose of supporting EGIC's activities, and we will ensure the management of generated AI and the detection of risks.
2. Scope and Operating Policy
Scope of Application: This policy applies to all EGIC employees, promotes the prohibition of unauthorized generation and plagiarism, and expresses the organization's policy.
Policy on the Use of Generated AI:
- Prohibition of Plagiarism: We will ensure that generated AI is not used based on other people's content.
- Clarification of Generated Content: Content that is primarily generated by AI will be clearly identified.
- Ethics and Transparency: Avoid unauthorized modification or imitation, and operate in a transparent manner.
3. Fact-checking and management
- Verification with multiple sources of trusted information: We conduct at least two or more fact-checks based on official information sources, such as the Japan Meteorological Agency and local governments, to ensure the accuracy of information.
- Deployment of fact-checking staff: Regardless of AI-generated content, we ensure thorough review by two or more people to prevent misinformation and inappropriate expressions.
- Human fact-checking: AI-generated content is always reviewed by human eyes to remove inaccurate information and confusing content.
4. Measures for the use of AI generated within the organization
- Clarify and limit the purpose of use: The use of generated AI within the organization should be limited to support tasks such as translation and proofreading, and should not be used for decision making or information sharing.
- Reporting requirements: The purpose and content of generated AI should be shared within the organization and reported to senior management in a timely manner (within 3 days).
- Education and Training: We will conduct regular training and education on the risks of AI generation and appropriate use, and confirm appropriate management and measures to prevent infringement.
5. Response to Violations
- Investigation and Corrective Action: If a violation is suspected, we will promptly investigate and correct the violation and, if necessary, suspend or delete the information.
- Guidance and Education to Prevent Recurrence: Provide ongoing training on rights and ethics.
- Establish an external reporting window: A window (support@ewrs.jp) will be established to receive external reports of unauthorized generation and plagiarism, and appropriate action will be taken.
6. promoting the use of generated AI and building trust
- Knowledge Dissemination: We will disseminate knowledge about the appropriate use of generated AI and promote ethical use.
- Provide reliable information: We will ensure the reliability of disaster prevention information and avoid inappropriate expressions and misunderstandings. We will not use generated AI to collect information.
- Improve collaboration and transparency: We will work with partner organizations and other institutions to share and disseminate information on the appropriate use of AI.
7.Revisions and Updates
EGIC will periodically revise these guidelines in response to technological developments and changes in risks, and will promote reliable information dissemination and the appropriate use of generated AI.
INQUIRIES
Questions regarding the Guidelines or reports of unauthorized generation or plagiarism should be directed to the official EGIC contact point (support@ewrs.jp).
Additional Information
These guidelines were established on November 9 and are effective November 14. The use of personal images of EGIC employees is permitted, provided that they do not violate the rights of others and are personal property.