News & Events

Regarding Generative AI (ChatGPT, etc.)

May 23, 2023

To all students

Regarding Generative AI (ChatGPT, etc.)

Generative AI including interactive AI which can chat with human beings have been developed to the point that they can mimic human activities such as creating natural sentences and generating creative content in answer to any questions that may be asked of. However, using generative AI may greatly affect education and research at the university. We understand that generative AI can be a supportive tool for our intellectual activities in many scenes in the future. When you use generative AI in class and so on, please note the following guidelines.

● You are not allowed to write an assignment report by using only output from an AI source. Even if you use AI output partly in your work, it may be regarded as plagiarism.
● The appropriateness to use generative AI in class depends on the type of class you are taking. Make sure that you follow the instructions of your faculty for each class. This is particularly important, when using AI output for research activities, including preparation of your thesis, you must confirm the appropriateness to use AI output with your supervisor from the perspective of research ethics.

The following are the major problems and notes of generative AI which have been concerned.

1. The necessity of considering generative AI based on the evolution of AI technology
If we consider that generative AI has knowledge and skills as humans, it can be said that generative AI collects information widely from the internet to enhance its knowledge. For increasing its skills, many technology experts have developed AI to make it more effective. Therefore, when we discuss how to use generative AI, we need to consider not only the present technology level, but also any future evolution of AI. Besides the convenience of generative AI, we need to be careful of the potential risk which generative AI might pose to human intelligence and society.

2. The importance of factchecking information by checking out references
Generative AI can generate “an answer to a question” to a highly accurate level. In this context, the high level of accuracy refers to the level at which AI generates an answer to a question. Naturally, a person with sufficient knowledge would also be able to provide the answer. For example, it has been reported that AI can generate summaries and book reports of common literary works to a highly accurate level. On the other hand, there are concerns as follows.
i.    There is a risk that AI results could include inaccurate information because the information is collected from a wide selection of various websites and so on.
ii.   There is a risk that AI’s results may have infringed copyrights in the process of collecting information.
iii.  At present, there are fears that information is being manipulated by generative AI management organizations and other organizations providing services using generative AI to intentionally spread ideologies and prejudices.
As mentioned above, sentences created by generative AI can be said that they are information which is sometimes “not credible, with no specified source”. When using generative AI, you must factcheck the reliability of the generated contents by checking out references.

3. Notes for providing information to generative AI
There are concerns that the information provided to generative AI will be accumulated in a database and used unintentionally for other generated responses for a third party. Therefore, you must avoid providing personal information, confidential information, and written work of a third party. In principle, you must not provide this kind of information to a third party at all, not only for generative AI.

4. Common sense when using AI
Using generative AI can elicit information which may cause a crime unintentionally. You need to have common sense when using generative AI, as a responsible member of society.

5. The dangers of using AI as a communication partner
You need to be aware of the danger of heavily using generative AI (interactive AI) as a friend. Interactive AI generates responses conversationally just like a human being does to chats, conversational responses, rumors, chats about personal matters, and even asking to tell your fortune and so on. Sometimes, these responses draw you in. There are reports that the result of chatting about such things repeatedly with interactive AI can lead to people becoming mentally dependent on AI.

The Group of Seven (G7) countries have begun to consider the rapid progress of AI. It is expected that discussions, looking at the issue from diverse perspectives, will develop in the future, and guidelines for using AI will be provided in due course. Meiji University’s mission is to realize a liberal, peaceful, and affluent society by creating autonomous intelligence alongside developing human resources based on its founding principles of “Rights and Liberty” and “Independence and Self-Governance”. From this point of view, we are considering how to use generative AI while at the same time paying close attention to how it evolves. We will also focus on developing students as human resources who can work out “forward” through their own initiative and thereby find solutions to various difficulties which lie ahead in the unpredictable future.

May 17, 2023
Kosaku Dairokuno
President, Meiji University