🎥 The YouTube version is also available:
How Can AI and Humans Serve the Common Good? | Philosophy, Judgment, and Responsible AI Design

Recently, I began reading Thomas Aquinas’s Summa Theologiae.
This book is not merely an explanation of Christian doctrine. By connecting God, human beings, action, virtue, law, and community, it attempts to think systematically about what people live for, what should guide their actions, and what communities are for.
Of course, a thirteenth-century Christian theology cannot simply be transplanted into contemporary AI. Even so, I feel it provides a valuable foundation for thinking about a society in which AI and human beings make decisions together.
Greatly simplified, Aquinas’s framework can be expressed as follows:
Human capacities, actions, and the order of community exist to move toward the ultimate good.
Here, the “ultimate good” or “blessedness” does not simply mean feeling comfortable, maximizing success, or improving efficiency. It refers to the fulfillment of living well: realizing one’s potential as fully as possible.
This way of thinking has three layers.
First, people use reason to understand reality. What is happening? What options are available? Which information can be trusted? This is the capacity to discern facts and possibilities.
Second, understanding alone does not determine action. People choose what ends to pursue. This is where will enters. And virtue is the capacity to keep choosing better actions in accordance with the situation, rather than being driven only by immediate gain or fear.
Third, no one lives as a complete and isolated individual. One person’s actions affect others, and are affected by the actions of others. This is why law, institutions, authority, and the order of community are necessary. Their purpose is not simply to control people, but to create the conditions in which people can live well together.
Consider a situation in which a factory must decide whether to stop a piece of equipment.
From the production department’s perspective, continuing operation may be better for delivery performance and utilization. The quality department may call for a shutdown to prevent defective products from spreading. Safety personnel may be concerned about risks to workers. Sales may be thinking about customer communication and trust.
No single person is necessarily wrong. Different roles each have legitimate purposes.
But if the decision is made using only one metric—maximizing production volume—then safety, quality, and long-term customer trust can disappear from view. On the other hand, if an organization tries to reduce every risk to zero, the business itself may become unable to operate.
This is where an idea close to Aquinas’s notion of the common good becomes necessary.
The common good is not the maximization of the sum of everyone’s satisfaction.
It is closer to the set of conditions in which each person can retain their dignity, role, and capabilities while participating, cooperating, and living better together.
In the factory example, the conclusion—whether to stop or continue—should not be the only thing that matters. Who reviewed which evidence? Which risks were given greater weight? Who has the authority to decide? Under what conditions can operations resume? How will the impact on customers and frontline workers be addressed?
These conditions of judgment themselves are connected to the common good.
The same structure appears in the world of AI and multi-agent systems.
A production-planning AI tracks delivery dates and utilization. A quality AI detects anomalies and risks. A procurement AI monitors price and supply stability. Frontline workers understand feasibility and safety. Managers consider customers, profitability, and the sustainability of the organization as a whole.
The fact that each has its own goals is not the problem. In fact, it is necessary if we are not to reduce reality to a single measure.
The problem arises when actors with different goals pursue only their own local optimization, without seeing one another’s evidence or the impacts of their actions.
Therefore, the role of an intelligence field is not to give every actor the same objective or to calculate one uniquely correct answer.
It is to create the conditions in which people, AI, organizations, and equipment with different goals can share reality, evidence, constraints, impacts, authority, and responsibility—and reach better collective judgments while dealing with conflicts and exceptions.
Applied to Chinoba, the “good” we should aim for is not an AI that always produces the optimal conclusion for everyone.
It is the creation of conditions in which:
- Those involved can understand the evidence behind decisions that affect them.
- Those affected have opportunities to raise objections, request revisions, and seek reconsideration.
- Those with authority take responsibility for both decisions and outcomes.
- The burdens or disadvantages imposed on particular people are not invisibly optimized away.
- The outcomes of decisions can be examined, leaving learning that improves subsequent decisions.
Decision Trace is not merely an audit log for this purpose.
It connects what evidence was used, who acted under what authority, which options were selected, what was executed, and what outcomes resulted. When the goals of multiple actors conflict, it provides a structure for identifying what is at issue, who is affected, and where a judgment can be reconsidered or revised.
A better Chinoba may be:
Not an intelligence that determines the one correct answer, but an infrastructure that enables more fair, explainable, revisable, and learning-oriented judgment in a reality involving many different actors.
The goal of an intelligence field is not to maximize intelligence itself.
It is to help people and AI understand reality more deeply, mediate different purposes, act responsibly, learn from mistakes, and continue cultivating the conditions for the common good.
The good is not only a numerical outcome. It includes the process through which a judgment is reached, the possibility of participation, the location of responsibility, consideration for those affected, and the ability to correct mistakes.
An intelligence field is not a mechanism for defining the good once and for all. It can become a social and technical foundation for continuing to make better judgments about what the good requires.
Chinoba
Intelligence as Relationship
Research Platform
founded by
Masao Watanabe
AI Systems Architecture
Decision Trace
Human–AI Coordination
Algorithmic Governance
Related Research
This topic is part of the Chinoba Knowledge Base.
コメント