As an international board committed to the ethical advancement of artificial intelligence, we endorse the following principles as the foundation for responsible AI development, deployment, and governance.

AI systems must be designed to enhance human agency, dignity, and well-being. Technology should support—not replace—human decision-making, especially in sensitive domains like healthcare, justice, and education.

AI systems must be understandable and auditable. Users and stakeholders should have access to meaningful explanations of how decisions are made, especially in high-stakes contexts.

• Clear documentation of data, models, and decision logic
   
• Avoidance of “black box” systems in critical areas
   

AI must be free from unjust bias and promote equity across race, gender, geography, ability, and socioeconomic status.

• Active bias detection and mitigation processes
   
• Inclusive data sourcing and validation
   
• Ongoing fairness audits
   

AI must respect individual privacy rights and adhere to data protection laws and best practices.

• Data minimization and anonymization by default
   
• Transparent consent processes
   
• Security protocols to prevent misuse or leakage
   

There must be clear accountability for the outcomes of AI systems.

• Defined roles and responsibilities across the AI lifecycle
   
• Redress mechanisms for affected individuals
   
• Internal governance policies and oversight boards
   

AI systems must be technically reliable and resilient to errors, manipulation, or adversarial attacks.

• Rigorous testing and validation
   
• Continuous monitoring and retraining
   
• Fail-safes and escalation paths in critical use cases
   

AI must be developed and deployed in ways that minimize ecological impact.

• Efficient training practices (e.g., energy-conscious models)
   
• Consideration of carbon footprint in scaling decisions
   

AI governance should be inclusive, culturally aware, and globally collaborative.

• Avoid technological colonialism
   
• Include perspectives from the Global South and underrepresented communities
   
• Align with global frameworks (e.g., UN SDGs, OECD AI Principles)
   

Above all, AI must not be used to cause harm, enable repression, or violate human rights.

• No development or deployment in support of mass surveillance, lethal autonomous weapons, or misinformation campaigns
   
• Proactive evaluation of unintended consequences
   

Ethics is not static. Organizations must continually learn, improve, and adapt their AI practices.

• Periodic ethics reviews
   
• Stakeholder feedback integration
   
• Alignment with emerging global standards