Zhu Consulting

In the international landscape including the UN system itself, there are multiple platforms advancing normative frameworks such as the ITU, UNESCO, the Global Digital Compact, EU AI Act, etc. In July, Geneva will also host multiple convenes on AI including AI for Good and WSIS. It will be important to distinguish the purpose of this platform from the others to avoid duplicating workstreams. This could be the following: AI is scaling rapidly - and so are its compounding risks. Government and international laws need to step in before AI is scaled into dangerous territories. Beyond normative consensus, we urgently need implementation: international and national governance/legal structures that translate value expectations such as rights and safety into concrete enforceable reality. This layer is absolutely essential. Until then, normative principles remain aspirational and vulnerable, largely dependent on the goodwill of individual actors, especially the few leading the trends. UN could facilitate in creating those international laws and support Member States in doing so, along with providing inclusive consultations and platforms for cross-boundary collaboration. Several high risks dimensions should be addressed, including preventing military misuse, limiting the development of superintelligence, promoting international cooperation, and other areas with potential existential risks for humanity. Beyond consensus, this platform should generate concrete governance outputs and structures to support Member States and international bodies to adopt robust mechanisms to implement normative expectations. This includes, for example, defining the UN's role in supporting governance at both international and national levels, mandating drafting of international laws, recommending scopes for national laws, setting implementation timeline, establishing mutual monitoring frameworks (including what happens in between annual dialogues), and creating mechanisms for cross-boundary collaboration. Making the convene as well as the outcome transparently available to the public would also be valuable. Continuous open consultations such as this are also extremely valuable and appreciated.

• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Interoperability of governance approaches
• Transparency, accountability, and human oversight
• Open-source software, open data and open AI models

While all are extremely relevant and important, the selected four especially reflect urgent priorities needed. This breaks down to two key aspects: 1) ensuring AI is aligned with humanity's collective purpose and wellbeing, and 2) ensuring AI governance is effective enough to implement value expectations in reality. Some topics are cross-cutting, for example, protecting and promoting human rights includes "social, economic, ethical, cultural, linguistic and technical implication of AI". I chose the second because it is more concrete layout of rights dimensions. "Safe, secure and trustworthy AI" entails "transparency, accountability, and human oversight" - perhaps they can be folded into one? Open source is important for enabling responsible AI in practice.

The Global Dialogue on AI Governance touches upon comprehensive core responsible AI and AI governance issues. Additional essential aspects not covered include necessary international and national governance frameworks including laws for value implementation. AI regulation urgently needs national and international laws to step in and regulate the industries. This would be essential to prevent AI scaling further into dangerous positions and become harder to address later on. More rationale please refer to answer to 8. Inclusion and cooperation across geographical regions could also be relevant.

Coming from the UN system previously, my work focuses on ensuring AI aligns with humanity's important values and supporting value implementation in concrete organizational practices across cultures, geographies, and social contexts (sociotechnical alignment, responsible stewardship, and positive innovation). While my practice is US-based, it is globally facing, supporting all sectors worldwide. The selected areas are repeatedly observed as challenges across contexts: scaling AI without proper consideration and understanding of its potential risks, users trusting AI without adequate digital literacy or safety boundaries, organizations lacking robust governance structures in place especially on AI ethics and safety, largely chaotic and scattered landscape to navigate AI trajectory, unemployment, infrastructure/capacity gaps, misinformation, and other rights issues such as privacy, exclusion, biases, and overriding human agency. Development of superintelligence, military misuse, atrophy of human talents, and national competitions at the cost of multilateral cooperation are also worrisome. While AI has tremendous opportunities, e.g. productivity, forecasting, analysis, research, translation, database monitoring, assistance in accessibility issues, etc., its benefits truly get unlocked if the technology is used right.

Please see answer to 8. Beyond consensus, this platform should generate concrete governance outputs and structures to support Member States and international bodies to adopt robust mechanisms to implement normative expectations. This includes, for example, defining the UN's role in supporting governance at both international and national levels, mandating drafting of international laws, recommending scopes for national laws, setting implementation timeline, establishing mutual monitoring frameworks (including what happens in between annual dialogues), and creating mechanisms for cross-boundary collaboration.

ITU AI for Good, UNESCO, WSIS, Global Digital Compact other implementing agencies, and major regional/national AI regulatory mechanisms e.g. EU AI act, US, and China.

Having open consultations such as these, having the sessions available for public watch, making results public, hosting hybrid sessions with possibility for virtual speaking participations engaging underrepresented communities.

Global South, work forces most vulnerable to being replaced by AI, civil society groups, women, people with disability, youth, and other populations at risk such as those without digital access and at poverty regions. Education sector should also be more included given the importance of children protection and protection of human talents.

Please see answer to 15.

UNESCO recommendations, Montreal Memorandum (April 2026), Global Digital Compact, ILO on labor rights and social justice, EU AI Act, the latest Chinese Generative AI Regulations. Helpful publications: UNDP Great Divergence Report, and UNU Unlocking AI's Potential to Serve Humanity.