We are committed to solving increasingly complex and sophisticated societal challenges and creating new value under our purpose of "Sustain the prosperity of society." We believe that artificial intelligence (AI) is a powerful key to realizing this purpose by dramatically augmenting human potential and creativity.

At the same time, the revolutionary benefits brought by AI must be built upon "Trust" and "Responsibility" in its development, deployment, and utilization. We consider the potential risks that may arise as AI becomes increasingly embedded in human society, and we recognize our social responsibility, more than ever, to uphold human dignity and human rights and to ensure fairness, transparency, and safety.

Rather than confining AI to being merely a tool for efficiency, we aim to position it as a trusted partner that enhances human well-being and contributes to building a sustainable future. To this end, we hereby establish the "Ethics Policy for Trustworthy and Responsible AI" as follows.

This Policy shall serve as a guiding principle for all of our business activities, and together with our stakeholders, we will contribute to realizing an ethical and transparent AI-enabled society.

1. Human-Centric

• Respect for Human Dignity and Fundamental Human Rights

  We position artificial intelligence (AI) as a means to augment human capabilities and contribute to social prosperity and the resolution of societal challenges, and adopt a human-centered approach as a fundamental principle. In all business activities, we respect human dignity and fundamental human rights and comply with international norms, applicable laws and regulations in each country, and our internal Human Rights Policy throughout all phases of the development, deployment, and utilization of AI.

• Consideration for Psychological Well-being

  We strive to ensure that the use of AI does not adversely affect human mental health or well-being. We prohibit practices such as unfair psychological manipulation (including dark patterns), exploitation of users' vulnerabilities, social scoring, and biometric identification without appropriate consent, and incorporate considerations for enhancing psychological well-being from the design phase. In addition, we consider the impact of AI adoption on human work and roles and seek to avoid creating excessive employment-related anxiety or undermining work motivation. By fostering new value and creativity that are uniquely human, we aim to ensure psychological safety across society, organizations, and individuals.

• Human Oversight and Control

  To prevent significant impacts on life, physical safety, property, and other important interests, we aim to ensure that human oversight and control can be exercised over AI-driven decisions as appropriate. We also ensure appropriate override mechanisms to prevent excessive autonomy of AI systems.

• Consideration for Users' Decision-Making

  We provide users with timely and appropriate information and take care not to impair their freedom of judgment, operation, or choice. Through the provision of interfaces that enable users to control AI in accordance with their own intentions, we support individual autonomy.

• Education and Training

  To realize this Policy, we will provide continuous education and training for all employees and other relevant stakeholders, with the aim of enhancing awareness and literacy regarding AI ethics.

2. Ensuring Diversity and Inclusion

• Promotion of Inclusive Design

  To extend the benefits of AI to all people, we will design AI systems and services with consideration for diverse users, including social minorities. Regardless of physical characteristics, cognitive abilities, language, or usage environment, we will strive to pursue universal design and improve accessibility so that everyone can access and benefit from our services without barriers.

• Reflection of Diverse Perspectives

  Throughout the entire AI development and evaluation process, we will strive to incorporate the perspectives of stakeholders with diverse backgrounds. By considering differences in culture, customs, and forms of expression across countries, regions, and languages, and by detecting and mitigating inappropriate biases arising from training data and algorithms biased toward particular values, we will work to prevent it in design.

3. Realization of a Sustainable Society

• Bridging the Digital Divide

  To ensure that AI-driven innovation does not create new inequality, we will provide fair and equitable opportunities through initiatives such as offering AI literacy education and developing user-friendly interfaces accessible to all.

• Reduction of Environmental Impacts

  Based on our Environmental Policy, we will promote the optimization of energy efficiency in AI model training and data center operations. In selecting technologies, we will prioritize options that minimize greenhouse gas emissions and reduce impacts on ecosystems.

4. Responsibility and Contribution to Society

• Contribution to Solving Social Issues

  We will maximize the potential of AI to realize our purpose of "Sustain the prosperity of society" and to contribute to the achievement of the Sustainable Development Goals (SDGs), thereby actively contributing to the resolution of growing social challenges.

• Compliance and Integrity

  We will establish a framework to ensure compliance with AI-related laws and regulations, and each employee will fulfill their responsibility to society with a strong sense of ethics and integrity.

• Protection and Respect for Rights

  In the development and operation of AI, we will operate strict management and verification processes to prevent the infringement of third-party rights, including copyrights, patent rights, and other intellectual property rights.

• Dialogue with Stakeholders

  We will promote transparent dialogue with diverse stakeholders and contribute to the development of a trustworthy AI society through collaboration and mutual cooperation among industry, government, and academia.

• Responsibility Across the Supply Chain

  We will also seek understanding of and support for the principles of this Policy from our business partners and suppliers involved in AI-related activities, and promote the realization of a responsible AI society throughout the entire supply chain.

5. Human Judgment, Involvement, and Controllability

• Effective oversight framework

  We will establish a framework to ensure appropriate human oversight throughout the entire AI lifecycle, from development to operation. In particular, where significant risks are foreseeable, humans will make the final decisions, and clear procedures and authority will be established to shutdown AI systems in response to unexpected behavior.

• Prohibition of unfair manipulation

  We will not develop or provide AI systems intended to unfairly manipulate human free decision-making or cognition at a subconscious level, nor will we engage in inducement that exploits human vulnerabilities.

6. Ensuring Safety and Security

• Ensuring physical and psychological safety

  We will implement thorough safety measures in the design of AI systems to prevent their use from causing not only physical harm, but also psychological distress or social disadvantage.

• Providing appropriate usage guidelines

  We will disclose the functions, limitations, and foreseeable risks of AI systems, and provide appropriate usage guidelines to prevent accidents caused by misuse or overreliance.

• Implementation of a risk-based approach

  We will conduct risk assessments based on the intended use of AI and implement appropriate safety measures and risk mitigation actions commensurate with the level of impacts. We will also carry out continuous monitoring after deployment.

• Management of data quality and avoidance of bias

  To ensure appropriate AI training, we will maintain the accuracy and comprehensiveness of data and operate management processes to identify and mitigate potential biases contained in training data.

• Cybersecurity and incident response

  To protect AI systems from cyberattacks and unauthorized use, we will implement robust security measures and ensure system resilience. In the event of an incident, we will work promptly with relevant stakeholders to prevent the spread of damage and recurrence.

• Ensuring interoperability

  We will strive to comply with international standards and enhance overall societal safety by ensuring interoperability and interconnectivity among different systems.

7. Protection of Privacy

• Strengthening data governance

  We will apply rigorous data governance to prevent data obtained through improper means or data of questionable quality from being used for AI training.

• Respect for privacy and legal compliance

  We will comply with our internal "Privacy Policy," "Personal Data Guiding Principles", and applicable laws and regulations, and appropriately protect and manage not only personal information but also all information related to privacy. Based on our existing personal information protection management framework, we will ensure transparency through the disclosure of the purposes of use, the consent acquisition from individuals, and the appropriate handling of requests for inquiries, suspension of use, and other related requests. In cases involving cross-border data transfers, we will assess the regulations and risks of the destination countries and implement appropriate legal measures. In addition, we will promote the concept of "Privacy by Design" to ensure privacy protection from the design phase onward.

8. Fairness and Non-Discrimination

• Prevention of unfair discrimination

  We will strive to ensure fairness so that AI does not lead to unfair discrimination or disadvantage based on specific attributes, including race, gender, sexual orientation, gender identity or expression, language, beliefs, religion, political opinion, social status, disability, age, or family relationship (including marital status).

• Detection and mitigation of inappropriate bias

  We recognize that bias may arise in AI decision-making and will technically and organizationally assess potential biases contained in training data, algorithms, and outputs.
  Where bias is identified, we will take appropriate measures to mitigate it.

9. Accountability and Transparency

• Pursuit of explainability

  We will strive to enhance the transparency of AI decision-making processes and promote the development of technologies that enable logical explanations of AI-generated outcomes (Explainable AI).

• Faithful disclosure of information

  We will provide users with sincere and clear explanations, within the bounds of technical and social reasonableness, regarding the use of AI, the methods for acquiring and utilizing data, and the appropriateness of AI-generated outcomes.

• Ensuring traceability

  We will appropriately retain records, including logs of data and processes that form the basis of AI decision-making, and ensure traceability that enables ex post verification.

• Clarification of accountability structures

  We will clarify roles and responsibilities in the development, deployment, and utilization of AI and maintain a governance framework capable of fulfilling accountability to society.

10. Administration of This Policy

This Policy shall be approved and overseen by the Board of Directors. We will review and revise this Policy as appropriate in response to changes in social conditions, technological developments, and laws and regulations.