2023.6
公立はこだて未来大学長 鈴木 恵二
ChatGPTやStable Diffusionなどに代表される生成系AIは、既存のテキストや画像などの大量のデータを学習して作られ、利用者の指示や質問(プロンプト)に応じてさまざまな種類のテキストや画像を柔軟に生成できるようになってきています。また、DeepLなどの翻訳AIも学習によって精度が向上し、実用的な翻訳文を出力可能になりつつあります。
生成系AI・翻訳AIは、まだ不完全ながらも、場合によっては高品質な出力を簡単に得られるため、世界中で急速に利用が進んでいます。しかし、生成系AIには剽窃や著作権侵害、個人情報の収集などさまざまな問題点があることも多くの人から指摘されています。
生成系AI・翻訳AIを授業や研究で適切に活用するために、本学では基本方針を以下のとおり定めます。
授業の課題(レポート)、筆記試験においては、以下のルールに従ってください。
ただし、授業や研究の目的・状況によって、生成系AI・翻訳AIの利用の仕方は異なります。それぞれの授業や研究において生成系AI・翻訳AIを利用するかどうかは、担当教員の指示に従ってください。また、担当教員から生成系AI・翻訳AI利用の具体的な指示やルール(クラスポリシー)が示された場合は、それに従ってください。
なお、ルールについては、AI技術の進歩に応じて定期的に見直しを行います。学部の卒業研究、大学院のシステム情報科学研究については、改めて通知します。
June, 2023
Future University Hakodate President SUZUKI, Keiji
Generative AI such as ChatGPT and Stable Diffusion is created by using large amounts of training data such as existing texts and images, and it is becoming more flexible at generating various types of texts and images in response to users’ instructions and questions (prompts). Moreover, AI-based translation services such as DeepL are also improving their accuracy through machine learning and becoming capable of outputting serviceable translations.
Generative AI and machine translation, though still imperfect, are being adopted rapidly around the world because they can, in some cases, easily produce high quality output. However, a variety of concerns have been raised including plagiarism, copyright infringement, and collection of personal information.
In order to make appropriate use of generative AI/machine translation in classes and research, Future University Hakodate has established the following basic policies.
– Use generative AI for your own learning purposes and take responsibility for the results of your use.
– Do not rely on the output of generative AI/machine translation as is; instead, examine the content carefully.
– Do not use the generative AI in a way that constitutes fraud such as plagiarism or fabrication.
– Be proactive and use the latest AI technologies in your research.
– Under the guidance of a faculty member, exploration of more effective learning methods using the latest AI technology is encouraged.
Please follow the guidelines below for class assignments (report) and written exam.
However, note that the use of generative AI/machine translation may differ depending on the purpose and situation of the class or research. Please follow the instructions of the faculty member as to whether or not to use generative AI/machine translation in each class and research. Moreover, if the faculty member gives you specific instructions or rules (class policy) for using generative AI/machine translation, please give them precedence over the guidelines summarized here.
– Basic liberal arts subjects and subjects taught in English also aim at improving language and writing skills. Therefore, in principle, the use of generative AI/machine translation is prohibited for the class assignments (reports) of these subjects.
– Use a citation and clearly indicate the source, even if you are allowed to use the generative AI/machine translation. The use of cut and paste of the generative AI/machine translation output into an assignment is prohibited.
– The user is responsible for checking that there are no problems with the output results of the generative AI/machine translation, and appropriately correcting if there are any problems.
– Use a reliable source of information to double check the output results of the generative AI/machine translation.
– The use of generative AI/machine translation is prohibited unless explicitly permitted by the faculty member.
These guidelines will be reviewed periodically in accordance with advances in generative AI technology. Graduation research at the undergraduate school and system information science research at the graduate school will be notified again.
Note:
This document was translated to English from the original written in Japanese, and its English content was checked by translating it back to Japanese using DeepL, and modified as necessary.