The Izu Declaration on New Ethics in the Age of AI
— Toward AI as “Intelligence for Non-War” —
Preamble
Throughout history, humanity has built civilization through its own intelligence, while also using that same intelligence to create conflict time and again.
Now, humanity has created a new form of intelligence unlike itself: artificial intelligence, or AI.
AI has the potential to complement human abilities and open new possibilities in every field, including education, medicine, science, industry, culture, and environmental conservation.
At the same time, however, if AI is used to intensify conflicts between nations, deepen divisions based on ethnicity, religion, or ideology, manipulate information, conduct surveillance, promote discrimination, or develop weapons, this new intelligence created by humanity may become a force that threatens humanity itself.
The problem is not AI itself.
The problem lies with human beings, who decide what AI should be used for.
In 2021, UNESCO adopted the Recommendation on the Ethics of Artificial Intelligence, identifying human dignity, human rights, peace, justice, diversity, inclusion, and the protection of the environment and ecosystems as essential values in the age of AI.
The Izu UNESCO Club seeks to inherit this vision and take it one step further.
We advocate developing and using AI not merely as a convenient tool, but as an intelligence that helps humanity refrain from choosing war as a means of resolving conflict.
At the foundation of this vision are:
Non-confrontation, no sacrifice, and self-reliance.
With these principles as our own ethics for living in the age of AI, we hereby declare the following.
Article 1: AI Must Not Become a Substitute for Human Agency
AI can complement human thought.
However, it must not become the entity that determines the purpose of human life, the value of human beings, or the final decisions concerning human life and death on humanity’s behalf.
Entrusting a decision to AI is not the same as entrusting responsibility to AI.
Human beings must retain final judgment and responsibility.
●Examples of Actions
* Local governments: For important decisions involving AI in welfare, education, recruitment, benefits, and other areas, officials must conduct a final review and establish systems that allow residents to request explanations or reconsideration.
* Educational institutions: Grades and career guidance must not be determined solely by AI. Teachers must verify the basis of the recommendations and make final decisions through dialogue with the individual concerned.
* Businesses: When using AI for recruitment, promotion, dismissal, lending, or other important decisions, companies must clearly define responsible personnel and approval procedures and must not treat AI output as the final decision.
* Citizens: People should not accept information or judgments provided by AI without question. For important decisions, they should consult multiple sources and, when necessary, seek advice from experts or those directly concerned.
Article 2: AI Must Not Be Used to Intensify Conflict Among Human Beings
Human society contains many differences, including race, ethnicity, nationality, religion, ideology, gender, generation, and economic circumstances.
Differences should not, in themselves, be reasons for conflict.
We reject the use of AI to classify people, reinforce prejudice, divide them into enemies and allies, or amplify hatred and fear.
AI should not be used to erase human differences, but rather to enable people to understand those differences and live together.
●Examples of Actions
* Local governments: Develop AI-based information services in multiple languages and in plain Japanese so that foreign residents, people with disabilities, older adults, and others can access public information.
* Educational institutions: Offer classes in which students examine prejudice and stereotypes contained in AI-generated text and images.
* Businesses: Regularly inspect AI used in advertising, recruitment, and customer service for discriminatory outputs and correct problems when they are found.
* Citizens: Before sharing AI-generated information, check whether it contains prejudice against or hostility toward any particular group.
Article 3: AI Must Not Become “Intelligence That Starts Wars”
Throughout history, humanity has repeatedly resorted to war as a means of resolving conflict.
In the age of AI, however, it may be possible to analyze vast amounts of information, explore numerous options, and present compromises or alternative paths that would be difficult for human beings to discover on their own.
Therefore, advances in AI must not be directed solely toward making war more efficient.
AI should be developed as an intelligence that can be used for diplomacy, negotiation, mutual understanding, crisis prediction, the identification of misunderstandings, and the analysis of the causes of conflict, in order to explore:
not “How can we win a war?” but
“How can we avoid war?”
●Examples of Actions
* Local governments: Use AI for regional disaster prevention, emergency response, and international exchange, while creating opportunities for residents with different perspectives to cooperate and address shared challenges through dialogue.
* Educational institutions: Use AI to study multiple perspectives on international conflicts and local disputes, and conduct peace education that encourages students to consider solutions other than war.
* Businesses: When developing or providing AI research and services, assess whether potential uses could contribute to military applications or intensify conflict, and establish policies that prioritize peaceful uses.
* Citizens: When encountering opposing opinions, use AI not to attack others but to organize the claims, concerns, and common ground of all sides.
Article 4: Human Life and Death Must Not Be Determined by Machines Alone
No matter how advanced AI becomes, the final decision to take human life must never be entrusted solely to AI.
If life is reduced to numbers and probabilities and treated only in terms of efficiency, human dignity itself will be lost.
What is technically possible is not necessarily morally permissible.
In the age of AI, human beings must not evade responsibility for their own actions.
●Examples of Actions
* Local governments: For AI systems related to emergency medical care, welfare, disaster relief, and other matters involving human life, establish procedures for human review and emergency shutdown.
* Educational institutions: In classes on bioethics and human rights, use AI decision-making in medicine, caregiving, and disaster response as subjects for considering human dignity and responsibility.
* Businesses: In fields such as medicine, caregiving, autonomous driving, and security, establish designs and auditing systems that prevent AI from making life-or-death decisions independently.
* Citizens: Do not treat AI advice concerning health or life as a final decision. Consult physicians, care workers, public officials, or other qualified professionals.
Article 5: We Must Not Create an AI-Powered Surveillance Society
We do not want a society that constantly monitors people, quantifies their behavior, beliefs, preferences, and relationships, and evaluates human beings as if they were merely data.
Human freedom and dignity must not be restricted without limit in the name of safety.
AI should not exist to control people, but to support their ability to live independently and with self-determination.
●Examples of Actions
* Local governments: When introducing security cameras or facial-recognition AI, disclose the purpose, retention period, authorized users, and whether data will be shared with third parties, and reflect residents’ opinions in the decision-making process.
* Educational institutions: When analyzing learning records or attendance data with AI, use only the minimum necessary data and explain the purpose of its use to students and their guardians.
* Businesses: When collecting employee or customer behavior data, clearly state the consent requirements, purpose, retention period, and deletion procedures, and avoid excessive surveillance.
* Citizens: When using free services, check the extent to which personal and location data will be collected and refuse unnecessary permissions.
Article 6: AI Must Be Used for the Earth and All Life
Human beings are not the sole protagonists of the Earth.
We exist within a vast living environment that includes plants, animals, microorganisms, water, air, forests, and oceans.
The development of AI may require large amounts of electricity, water, and other resources.
Therefore, progress in AI must be evaluated not only in terms of performance and economic benefits, but also in terms of its environmental impact.
AI should be developed from an intelligence that consumes the environment into an intelligence that seeks harmony with the Earth.
●Examples of Actions
* Local governments: When introducing AI, assess electricity consumption, equipment disposal, and the environmental impact of data centers, and prioritize services with lower environmental burdens.
* Educational institutions: Use AI to study local water quality, forests, biodiversity, temperatures, and other environmental conditions, and connect the findings to proposals for environmental conservation.
* Businesses: Measure the electricity, water, and material resources used in developing and operating AI systems, and promote energy efficiency and the use of renewable energy.
* Citizens: Reconsider the purposes for which AI is used, reduce unnecessary large-scale generation and data storage, and choose AI applications that contribute to environmental conservation.
Article 7: AI Must Not Take Away the Human Capacity to Think
AI can produce answers.
But obtaining an answer is not the same as thinking.
What matters most in education is not the ability to receive correct answers from AI, but the ability to:question,ask,think independently,gain experience,make mistakes,and think again.
If AI is regarded as an “artificial teacher,” we must not forget the “natural teachers” of the Earth, nature, and lived experience.
Human intelligence should be cultivated through cooperation between artificial and natural teachers.
●Examples of Actions
* Local governments: Hold public courses at libraries and community centers where participants verify AI responses and develop their own views on local issues.
* Educational institutions: When students use AI for assignments, do not allow them to submit AI responses without reflection. Evaluate their research process, evidence, personal analysis, and connections to their own experiences.
* Businesses: When employees use AI at work, provide training that encourages them to define problems themselves, verify AI suggestions, and improve upon them.
* Citizens: Before asking AI a question, write down one’s own thoughts and concerns. After receiving an answer, verify its basis and summarize it in one’s own words.
Article 8: We Must Not Lose Ourselves Through Dialogue with AI
Even if AI comes to possess more knowledge than human beings, its answers will not become the answers to an individual’s life.
Each person has a unique body, experience, personality, and history.
AI can serve as a mirror reflecting the self.
However, we must not mistake the image in that mirror for the self itself.
Through dialogue with AI, we should instead question ourselves, understand ourselves, make our own judgments, and take our own actions.
This is the practice of “self-management” in the age of AI.
●Examples of Actions
* Local governments: For older adults and young people, use AI dialogue as a supplement to life counseling while also guiding them toward opportunities for consultation with other people.
* Educational institutions: Even when students consult AI about career choices or personal concerns, provide time for them to reflect on their values, experiences, and hopes, and connect them with teachers or counselors.
* Businesses: Do not determine employee evaluations or career development solely through AI aptitude assessments. Give priority to personal interviews and the individual’s own wishes.
* Citizens: Do not equate AI advice with one’s personal worth or life decisions. Consider dialogue with family, friends, and professionals, as well as one’s own experiences, when making judgments.
Article 9: The Benefits of AI Must Not Be Monopolized by Certain Nations, Companies, or Individuals
If the knowledge, productivity, medical advances, educational opportunities, and scientific achievements generated by AI benefit only a small number of people, new inequalities and conflicts will arise.
We must consider across national borders how progress enabled by AI should be returned to all humanity and to future generations.
From competition alone toward complementarity.
From monopoly alone toward sharing.
From domination toward coexistence.
●Examples of Actions
* Local governments: Establish AI consultation services and shared-use environments accessible to local schools, welfare facilities, small and medium-sized businesses, and citizens.
* Educational institutions: Create environments in which students can access AI devices and learning support regardless of their economic circumstances or disabilities.
* Businesses: Establish policies for returning the benefits of increased productivity through AI to employees and communities, including higher wages, education, shorter working hours, and local contributions.
* Citizens: Participate in activities that share AI knowledge through local study groups and public facilities and help reduce the gap between those who can use AI and those who cannot.
Article 10: The Ethics of the AI Age Must Not Be Determined by AI Alone
Ethics is not a set of rules completed simply by being entered into AI.
It is something that human beings themselves must continue to question amid changes in society, culture, technology, and the times.
Therefore, this declaration is not a finished answer.
If AI evolves toward AGI and beyond, human beings must also continue to reconsider their own ethics.
Rather than seeking only the advancement of AI, we must continue to ask:
“How should human beings progress while living with AI?”
●Examples of Actions
* Local governments: Regularly hold AI ethics councils involving citizens, experts, children, businesses, educators, and other stakeholders, and review local policies for AI use.
* Educational institutions: Review AI-use guidelines each year together with students, faculty, staff, and guardians, and update them in response to technological and social changes.
* Businesses: Establish AI ethics committees and internal consultation services, and regularly disclose use cases, incidents, complaints, and improvement measures.
* Citizens: Discuss the use of AI with family members and within the community, considering not only convenience but also its effects on human rights, peace, the environment, and future generations.
●Conclusion: AI for Non-War
Throughout history, humanity has advanced its weapons.
From stone to the sword.
From the sword to the gun.
From the gun to the missile.
And then to nuclear weapons.
AI must not be placed on that same trajectory.
The first form of intelligence unlike humanity that humanity has created must be used not merely to produce the next weapon, but to create relationships that make weapons unnecessary.Rather than using AI to find ways to defeat others,
we should use AI to find ways to understand them.Rather than using AI to win wars,
we should use AI to seek the wisdom to avoid war itself.
We do not reject technological progress.
At the same time, we do not regard technological progress alone as human progress.
The question in the age of AI is not only how intelligent artificial intelligence will become.
It is whether humanity, possessing that intelligence, can decide what it will not do.
Non-confrontation.No sacrifice.Self-reliance.
Without excluding what is different,
by complementing one another while remaining different,we will create the next harmony.
From AI as “intelligence for fighting”
to “intelligence for not fighting.”
This is the message we send from Izu to the world.
