Self Employed

I hope the following three key outcomes can be achieved within 12 months: • Set shared principles: Agree on the basics. A binding, shared understanding of what good AI governance looks like: transparent, accountable, and designed to benefit people. No vague promises. • Set practical rules and share what works: A clear plan for developing global technical standards that anyone can follow, plus real-world best practices for building AI that people can trust. These should be actionable, not just theoretical. • Include everyone and act fast: A working framework that brings together governments, international organisations, companies, civil society, and tech communities. It must actively promote diversity in AI and include specific deadlines, checkpoints and review processes to ensure commitments turn into action within one year.

• Safe, secure and trustworthy AI
• AI capacity-building
• Interoperability of governance approaches
• Transparency, accountability, and human oversight

• Safe, secure and trustworthy AI: Without safety and security, people won't trust AI. This focuses on ensuring the AI ecosystem is built to protect human users, prevent harm and earn trust through reliability. • AI capacity-building: Empower people to build skills and ethical understanding of how to use AI tools effectively and responsibly. • Interoperability of governance approaches: Focus on establishing structures and decision-making processes for AI data exchange and integration, AI and data interoperability. This involves establishing clear duties and authorities for government institutions while creating coordination mechanisms that engage the private sector, civil society organisations, academia and other relevant stakeholders for better governance. • Transparency, accountability, and human oversight: Encourage transparency and the responsible use of AI and data to foster innovation by making systems open to scrutiny, holding outcomes accountable and keeping humans in the loop. This is to maintain a safe, secure and trustworthy AI ecosystem.

• Continuously enhance AI and data infrastructure: Invest in resources such as AI data repositories, integration platforms and analytical tools to support data management, analysis and sharing. Develop technology roadmaps to identify specific needs and assess the current state of the technology infrastructure. • Build an empowered, diverse team to ensure AI diversity: Bring together individuals with expertise from different cultural and linguistic backgrounds, working domains to leverage collective skills and knowledge to make informed decisions and foster innovative ways of thinking and working. • Deploy human-centred design principles to design AI models that prioritise human rights, privacy, safety: Design AI models with a human-centred mindset to help solve real problems and meet human needs rather than prioritise attention, engagement and commercial/political profits.

Challenges: 1. Poor data quality for building AI models 2. Inconsistent processes and no shared rules or standards 3. Lack of interoperability, no integrated development environment for better integration and efficiency 4. Lack of an open and innovative culture to help build people's capabilities in using AI Opportunities: 1. Leverage open data and AI models and strengthen data access to allow civil society and academia to operate freely and safely 2. Set basic rules, core standards and policies for AI use and governance 3. Build stronger partnerships with trustworthy AI industries and invest in AI and data infrastructure 4. Promote a safe, open and learning environment to help build and improve people's AI literacy and skills.

Serve as an inclusive, safe and trusted platform where diverse groups can openly share knowledge, align expectations, and learn from one another.

National government policies like Singapore National AI Strategy and international AI principles or standards like EU AI Act, OECD AI Principles. UN AI Global Dialogue would bring a larger and more diverse discussion forum for better collaboration and impact if it were well-facilitated.