Keio University

Takumi Shimizu: Intellectual Production in the Era of Generative AI

Writer Profile

  • Takumi Shimizu

    Faculty of Policy Management Associate Professor

    Specialization: Management Information Systems

    Takumi Shimizu

    Faculty of Policy Management Associate Professor

    Specialization: Management Information Systems

2024/07/22

This article is co-authored with generative AI. I research the co-evolution of technology and organizations, and I am interested in how generative AI changes the intellectual production of individuals and organizations. I believe that generative AI has the potential to play a role as an interface for human-machine collaboration, going beyond its role as a convenient tool for assisting and automating tasks.

Recently, I read an article about the use of generative AI in business by an executive at a global company. The article discussed how the executive mastered generative AI as a tool, but to me, it felt as though they were collaborating with generative AI almost on the same level as a colleague or secretary (i.e., a human coworker). Regardless of the subjective view of the person involved, I believe this suggests that, as an objective reality, an era is arriving where we engage in intellectual production by viewing humans, machines, and algorithms in parallel.

To experience that potential firsthand, I attempted to write this article by editing text output by generative AI (ChatGPT). However, as a result, I ended up rewriting almost every sentence from scratch. ChatGPT suggested several important topics and provided a flawless draft that considered both positive and negative aspects, but using that as a base made the writing superficial and failed to convey my own voice effectively. Among those around me, many who take the integration of generative AI into their own intellectual production seriously are the ones pointing out its limitations.

However, this does not necessarily mean that the quality of generative AI is low. In fact, I use generative AI in various situations, such as coding and English proofreading, and it leads to faster and more accurate work than if I were to create it from scratch. It can be used extremely effectively for tasks where there is a general right or wrong answer or a clear standard of quality, and where I can confirm and verify it myself. On the other hand, when it comes to tasks like essay writing, I feel it is quite difficult to get text that can be used as-is beyond just brainstorming themes. I realized that even more than the further evolution of generative AI, there is a need for the author (the human side) to evolve the way they collaborate with it. The hurdle for AI to permeate widely into work, organizations, and management in the future may lie on the human side rather than the machine side.

*Affiliations and titles are as of the time of publication.