This course supports the Mission Statement of Nagoya University of Commerce & Business by developing leaders who can hold an organization accountable for what its systems do. Participants sit in the seats of directors and executives who must decide how far to delegate judgment to machines, when to deploy a system that cannot be fully verified, and who answers when harm follows. Cases span the United States, Europe, India, and Japan, and the regulatory divergence among them is itself part of the material, so participants practice reasoning across the boundary between the new era of Asia and the rest of the world. In requiring participants to take ethical positions and defend them under challenge, the course cultivates the Frontier Spirit and the ethical leadership the mission calls for.

Boards and executives are being asked to govern systems they did not build and cannot fully explain. The classical instruments of corporate governance — the duty of oversight, the allocation of decision rights, disclosure, and legal liability — were designed for a firm in which judgment sat with people. Artificial intelligence moves some of that judgment into systems that act at scale, fail in correlated ways, and cannot account for themselves.

By studying this course, participants will deepen their understanding of what changes and what does not. The course begins deliberately with governance failures that involved no AI at all, because the argument that AI makes oversight harder can only be made against a clear picture of how oversight already fails. It then works through the decisions AI actually forces: whether to codify expert judgment, when to release an unverifiable system, how far to delegate authority, who is accountable for autonomous action, what must be disclosed, and where a firm should draw a line the law does not yet require.

The course is taught entirely through the case method. There are no lectures in the conventional sense. Each session centers on a real organization facing a real decision, and participants take the position of those who had to decide. Class time is spent building an argument collectively: participants take positions, defend them under questioning, and revise them when a colleague's reasoning proves stronger. Frameworks and concepts emerge from the discussion rather than preceding it.

By the end of this course, participants will be able to,

Knowledge and Understanding: Explain the separation of ownership and control, the agency problem, the duty of care and the duty of oversight, and how these apply — and fail to apply — when judgment is exercised by a system,
Specialized Skills: Assess whether an organization's governance and control structures are adequate to the AI it deploys; judge when a system is ready to release; allocate decision rights among board, management and machine; trace accountability across developer, deployer, executive and director,
General Abilities: Analyze an unstructured situation, identify the decision that must actually be made, and take a defensible ethical position under incomplete information; argue persuasively in a demanding discussion environment,
Attitudes: Recognize responsibility as something an organization holds rather than something it can transfer to a vendor or a system, and weigh competitive cost against restraint honestly.

*本学の教育ミッションを具現化する形で設定されています。

By completing this course, participants will gain,

The ability to read an organization's governance structure and identify where oversight will fail before it does,
A framework for deciding whether and how far to codify expert judgment into a system, and for anticipating what becomes fragile when scale is achieved,
The judgment to decide whether to deploy a system that cannot be fully verified, and to set the exchange rate between safety and speed deliberately rather than by default,
The ability to allocate decision rights among boards, managers and autonomous systems, and to specify what supervision of an agent actually consists of,
A method for tracing accountability across developer, deployer, executive and director when harm occurs,
An understanding of what must be disclosed, to whom, and when — including to regulators — as AI is embedded in core processes,
The ability to judge where a firm should draw a line the law does not require, and to argue for it in competitive terms.

教育手法 Teaching Method		% of Course Time
インプット型 Traditional		0 %
参加者中心型 Participant-Centered Learning	ケースメソッド Case Method	100 %
	フィールドメソッド Field Method	0 %
合計 Total		100 %

I. Prerequisites for the Course

Participants are expected to dedicate 2 to 4 hours of preparation before each case. Three cases are discussed each day.

Pre-Class Preparation
Every case is distributed in advance. No case materials are handed out during class. Where a case comes in several parts — (A), (B), (C) — all parts are assigned and read together as one package before the session.
Read the assigned case in advance and come prepared to take a position and defend it.
Cold calls are a standard and graded part of this classroom.
Post-Class Review
Revisit the frameworks that emerged in discussion and apply them to an organization of your own choosing, in preparation for the final report.

Assignment questions are issued for every case in a separate assignment sheet. A laptop is required.

II. Final Report

Case to be used: all cases. Assignment: synthesize what you have learned across the cases discussed in class and design an AI governance structure. Detailed instructions will be announced during the course. Individually written and individually graded; participants may research and discuss together, but every submission is the participant's own work. Submission Deadline: 11:00 pm on the final class day. Submission Method: via Google Classroom (within 20 A4 pages, Times New Roman 12pt, double-spaced, PDF).

III. Feedback

There is no written preparatory report in this course. Feedback is given in three ways. First, contribution is acknowledged and challenged in class as discussion proceeds — this is the primary feedback mechanism of the case method, and it is continuous. Second, each participant receives an individual written comment on the quality of their class contribution at the end of the second day, so that there is time to act on it before the course concludes. Third, the final report receives written comments through Google Classroom on submission. No work is due after the course ends.

*成績は下記該当項目を基に決定されます。
*クラス貢献度合計はコールドコールと授業内での挙手発言の合算値です。

講師用内規準拠 Method of Assessment	Weights
コールドコール Cold Call	0 %
授業内での挙手発言 Class Contribution	70 %
クラス貢献度合計 Class Contribution Total	70 %
予習レポート Preparation Report	0 %
小テスト Quizzes / Tests	0 %
シミュレーション成績 Simulation	0 %
ケース試験 Case Exam	0 %
最終レポート Final Report	30 %
期末試験 Final Exam	0 %
参加者による相互評価 Peer Assessment	0 %
合計 Total	100 %

評価の留意事項　Notes on Evaluation Criteria

Participants are expected to study the three cases assigned for each day.

By university policy, active class participation is a fundamental component of this course, accounting for 70% of the final grade. In assessing class participation, emphasis will be placed on the quality of contributions rather than mere frequency. Thoughtful engagement in discussions — demonstrating logical reasoning and deep insights — will be valued over simply sharing facts, paraphrasing, or summarizing previous points.

This course deals with questions on which reasonable people disagree. Participants are expected to take positions and defend them, and are credited for changing a position when a colleague's reasoning warrants it. Declining to take a position is not neutrality.

Cold calls are a standard and graded part of this classroom. Participants who do not participate in discussion cannot learn by this method, and their grade will reflect that.

The final report is graded on the strength of original reasoning, not on summary. It is important to deliver your own thoughts and critical reasoning, built on the cases discussed in class.

Turning in the assignment with imperfect English is fine. Using generative AI to write your report is strictly prohibited; where I judge that a report has been written by AI, it will receive zero points.

All grading is individual. Participants may research and discuss together, but every submission is the participant's own work.

• 配布資料

N/A

Newly Assigned Course