Huawei

A successful first Global Dialogue on AI Governance would be defined less by grand declarations and more by tangible alignment and credible next steps. First, success would mean establishing a shared baseline: agreement on core principles such as safety, transparency, accountability, and human oversight. Not every country will align perfectly, but even partial consensus—especially among major AI-developing nations—would reduce fragmentation and signal a willingness to cooperate. Second, it should produce actionable outcomes rather than vague statements. This could include a roadmap for interoperable regulatory frameworks, commitments to share safety research, or the creation of working groups focused on urgent risks like misuse, bias, and frontier model governance. If participants leave knowing exactly what happens next and who is responsible, that's a strong indicator of success. Third, meaningful inclusion matters. A credible dialogue must amplify voices beyond major powers—particularly from the Global South, civil society, and technical experts. AI governance will only be legitimate if it reflects diverse economic, cultural, and political realities. Fourth, trust-building is critical. Even modest transparency measures—such as voluntary disclosures about advanced AI systems or agreement on evaluation standards—can reduce suspicion and prevent a race-to-the-bottom dynamic. Finally, success would be measured by continuity. If the dialogue establishes itself as an ongoing process with clear milestones, rather than a one-off event, it can evolve alongside the technology it seeks to govern. In short, alignment, actionability, inclusivity, trust, and continuity would together mark a genuinely successful first step.

• Safe, secure and trustworthy AI
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Protection and promotion of human rights

AI safety as an essential principle in the lifecycle of the technologies. Protection of fundamental rights as an overall rule.

No.

On fundamental rights, the main gap lies in uneven enforcement and interpretation of existing protections. Frameworks like the General Data Protection Regulation and the newer EU AI Act set strong standards on privacy, non-discrimination, and transparency. However, rapid AI deployment—especially in hiring, credit scoring, and policing—often outpaces regulatory clarity. This creates risks of algorithmic bias, opaque decision-making, and limited avenues for redress. For individuals, that can mean decisions affecting livelihoods or access to services without meaningful explanation or accountability. For businesses, it introduces legal uncertainty and compliance costs, particularly when operating across multiple jurisdictions. At the same time, governance advances are pushing improvements. The classification of "high-risk" AI systems under the AI Act and requirements for explainability are encouraging companies to adopt better auditing, documentation, and human oversight practices. Civil society and courts are also becoming more active in testing these rights in practice. On AI safety, the gap is even more pronounced at the frontier level. While Europe emphasizes "trustworthy AI," there is still limited global alignment on evaluating and mitigating risks from highly capable systems (e.g., systemic failures, misuse, or dual-use capabilities). This affects both public institutions and private firms: governments must prepare for misinformation, cybersecurity threats, and critical infrastructure risks, while companies face pressure to demonstrate robust safety measures without clear, harmonized standards. Recent developments—such as increased investment in safety research, model evaluation benchmarks, and international coordination efforts—are positive, but fragmented. Without stronger cross-border cooperation, there is a risk of regulatory divergence or a "race between regimes," which could weaken both safety and rights protections.

The AI Dialogue can serve as a practical bridge between fragmented national approaches and the need for coordinated global governance. First, it can align expectations by fostering convergence on baseline principles—such as safety testing, risk classification, and respect for fundamental rights. Even if legal systems differ, agreeing on common definitions (e.g., what counts as "high-risk" AI) would reduce regulatory fragmentation and compliance burdens. Second, the Dialogue can act as a coordination hub for standards and technical practices. By bringing together governments, researchers, and industry, it can promote interoperable evaluation methods, auditing tools, and incident reporting mechanisms. This is especially important for frontier AI safety, where risks are cross-border by nature and unilateral regulation is insufficient. Third, it can enable confidence-building measures. Voluntary transparency commitments—such as sharing information about advanced AI capabilities, safety testing protocols, or major incidents—can reduce mistrust and the risk of competitive escalation. Over time, these soft commitments can evolve into more formal agreements. Fourth, the Dialogue can amplify underrepresented voices, particularly from developing countries that are often rule-takers rather than rule-makers. Inclusive participation would improve legitimacy and ensure governance frameworks reflect diverse economic and societal contexts. Fifth, it can support capacity building and knowledge sharing. Many countries lack the technical or institutional resources to implement effective AI governance. Coordinated training, shared research, and access to expertise can narrow this gap. Finally, its value depends on continuity. If structured as an ongoing process with working groups, milestones, and accountability mechanisms, the Dialogue can move beyond discussion toward implementation.

Partnerships with companies, tech associations must be considered

Open, professional and online format.

None.

Workshops and round tables.

AI risk regulation and data protection, transparency tools, etc