Arbitulis.AI

From a legal-entrepreneurial perspective, the success of the first Global Dialogue on AI Governance would be reflected in outcomes that reduce regulatory uncertainty while enabling responsible innovation. First, a successful Dialogue should lead to clarity in legal standards governing AI particularly on liability, due diligence, and compliance obligations. For a startup operating at the intersection of AI and arbitration, predictable rules on accountability and admissibility of AI-generated outputs would significantly enhance trust and adoption. Secondly, the development of harmonised or interoperable frameworks across jurisdictions would be critical. Fragmented regulation increases compliance costs and limits scalability. A unified baseline especially on data governance, confidentiality, and cross-border dispute resolution would allow legal-tech startups to expand globally with greater certainty. Thirdly, the recognition of AI tools within dispute resolution ecosystems would be a meaningful outcome. If the Dialogue acknowledges and encourages the integration of AI in arbitration, mediation, and ODR processes subject to safeguards it would legitimise and accelerate innovation in this sector. Finally, the creation of regulatory sandboxes and public-private collaboration mechanisms would be essential. Startups require safe environments to test AI systems in compliance with evolving legal standards, while also contributing to policy development. In essence, the Dialogue would be successful if it translates international legal principles into practical, innovation-enabling frameworks ensuring that emerging AI startups can operate with legal certainty, cross-border scalability, and institutional legitimacy.

• Safe, secure and trustworthy AI
• Protection and promotion of human rights
• Transparency, accountability, and human oversight
• Open-source software, open data and open AI models

My selections reflect a focus on building AI systems that can be meaningfully integrated into legal processes without compromising core legal principles. Safe and trustworthy AI is essential because any technology used in dispute resolution must inspire confidence and ensure procedural integrity. If AI systems are unreliable or vulnerable, their outputs cannot be relied upon in adjudicatory settings. The emphasis on human rights ensures that AI deployment remains consistent with established international obligations. In legal contexts, this is particularly important to safeguard fairness, equality before law, and access to justice. Transparency, accountability, and human oversight are critical to preserving due process. Legal outcomes must remain explainable and reviewable, and ultimate decision-making authority must rest with human actors to prevent arbitrariness. Finally, open-source and open AI frameworks support accessibility and wider participation. They enable the development of cost-effective, standardised solutions, which is particularly relevant for expanding access to dispute resolution mechanisms across jurisdictions. Taken together, these priorities aim to ensure that AI evolves as a tool that strengthens rather than undermines the legitimacy, accessibility, and efficiency of legal systems.

Yes. While the listed themes capture core governance concerns, certain cross-cutting issues merit distinct attention. First, liability and attribution in AI-assisted decision-making remains underdeveloped. Existing frameworks do not adequately address who bears responsibility when harm arises from semi-autonomous systems particularly in complex, multi-actor environments involving developers, deployers, and end-users. Secondly, evidentiary standards for AI-generated outputs require clearer articulation. In legal and quasi-judicial settings, questions of admissibility, probative value, and reliability of AI-assisted analysis are likely to become central, yet remain insufficiently addressed in current governance discourse. Thirdly, the issue of algorithmic opacity in proprietary systems poses a structural challenge. While transparency is recognised as a principle, practical access to underlying models, especially in commercially sensitive contexts, remains limited raising concerns for auditability and due process. Fourthly, regulatory fragmentation and conflict of laws is an emerging concern. Divergent national approaches to AI governance risk creating compliance burdens and legal uncertainty, particularly for cross-border digital services. Finally, long-term institutional adaptation is often overlooked. Legal systems, including courts and arbitral institutions, will need procedural reforms, capacity-building, and technological integration to effectively engage with AI-driven processes. These issues cut across existing themes and are critical to ensuring that AI governance evolves in a manner that is not only principled, but also legally operational and future-ready.

In the legal and dispute resolution sector, particularly in India and comparable jurisdictions, the identified governance gaps are creating both structural challenges and significant opportunities. A primary challenge lies in the absence of clear regulatory standards for AI deployment in legal processes. While AI tools are increasingly used for research, case management, and decision-support, there is limited guidance on their permissible scope, evidentiary value, or liability in case of error. This creates uncertainty for adoption within courts, arbitral institutions, and private platforms. Secondly, data governance and confidentiality concerns are especially acute in arbitration, which is premised on privacy. The use of AI systems raises questions regarding data storage, cross-border data transfers, and compliance with evolving data protection regimes, thereby affecting party trust. Thirdly, limited institutional capacity and digital infrastructure constrain effective integration. Many legal systems, particularly at the sub-national level, lack the technical expertise and resources to assess, deploy, or regulate AI tools. However, these gaps also present substantial opportunities. There is a growing scope for AI-enabled dispute resolution platforms to enhance efficiency, reduce costs, and improve access to justice particularly in high-volume, low-value disputes. Further, the absence of rigid regulation allows room for innovation through regulatory sandboxes and pilot frameworks, enabling responsible experimentation with AI in legal contexts. Finally, the increasing global focus on trustworthy AI creates an opportunity for jurisdictions and sectoral actors to shape emerging standards, positioning themselves as leaders in ethical and legally compliant AI adoption. Overall, the current landscape reflects a transition phase where governance gaps pose risks, but also enable transformative innovation in dispute resolution ecosystems.

The AI Dialogue can play a pivotal role in advancing international cooperation by acting as a norm-setting and coordination platform within the UN system. First, it can facilitate the development of shared baseline principles grounded in international law, particularly human rights and due process. A common normative framework would reduce regulatory fragmentation and provide States with a reference point for domestic legislation. Secondly, the Dialogue can promote regulatory interoperability by encouraging alignment across jurisdictions. This is critical for cross-border AI applications, including digital services and dispute resolution platforms, where conflicting legal regimes create uncertainty and compliance burdens. Thirdly, it can serve as a forum for structured knowledge-sharing and capacity-building, especially for developing countries. Bridging technical and regulatory gaps would ensure more inclusive participation in AI governance and prevent concentration of influence among a few advanced economies. Fourthly, the Dialogue can support the creation of institutional mechanisms, such as expert working groups or model frameworks, to continuously address emerging issues like liability, auditability, and data governance. Finally, it can encourage public–private collaboration, enabling engagement between States, industry, and academia. This is essential to ensure that governance frameworks remain practical, innovation-friendly, and responsive to technological developments. In essence, the AI Dialogue can move international governance from fragmented, reactive approaches toward a coherent, cooperative, and forward-looking legal order for AI.

The AI Dialogue should build upon existing multilateral, regional, and sector-specific initiatives to avoid duplication and to consolidate emerging norms into a coherent global framework. At the international level, instruments such as UNESCO's Recommendation on the Ethics of AI (2021) and the OECD AI Principles provide an established normative foundation centred on human rights, accountability, and transparency. Regionally, frameworks like the European Union's AI Act and the Council of Europe's work on a Framework Convention on AI offer advanced regulatory models that can inform global standards. In parallel, technical and multi-stakeholder initiatives—such as the Global Partnership on AI (GPAI) and standard-setting work by bodies like ISO/IEC—contribute operational and technical guidance. The Dialogue can add value by functioning as a convergence platform that connects these fragmented efforts. First, it can promote norm harmonisation, translating diverse principles into a more universally accepted baseline aligned with international law. Secondly, it can bridge the gap between high-level principles and practical implementation, particularly by facilitating model laws, guidelines, or best practices adaptable across jurisdictions. Further, the Dialogue can ensure inclusive participation, bringing in perspectives from developing countries that are often underrepresented in existing frameworks. This would enhance legitimacy and equity in global AI governance. Finally, it can enable cross-sector integration, linking legal, technical, and policy communities to address complex issues such as liability, auditability, and cross-border enforcement. In essence, the AI Dialogue's added value lies in transforming dispersed initiatives into a coordinated, inclusive, and implementation-oriented global governance architecture.

Effective participation in the AI Dialogue requires a multi-stakeholder structure where each actor contributes according to its institutional competence, while ensuring that outcomes remain legally grounded and implementable. States should lead in articulating national priorities, sharing regulatory experiences, and negotiating common principles. Their role is central in translating Dialogue outcomes into domestic legal frameworks. Industry and startups should contribute practical insights on deployment, risk management, and compliance challenges. Their participation ensures that governance frameworks remain innovation-compatible and technically feasible. Academia and legal experts should provide doctrinal clarity—particularly on issues such as liability, jurisdiction, and human rights implications—helping bridge the gap between technology and legal regulation. Civil society should represent public interest concerns, including rights protection, accessibility, and ethical safeguards, ensuring that governance remains inclusive and legitimate. From a structural perspective, the Dialogue should adopt a tiered and continuous format: 1. Plenary sessions for high-level consensus on principles and priorities. 2. Thematic working groups (e.g., liability, data governance, dispute resolution) tasked with developing detailed recommendations. 3. Technical and legal expert tracks to translate principles into model frameworks or guidelines. 4. Regulatory sandbox or pilot initiatives to test proposed approaches in controlled environments. Additionally, the Dialogue should function as an ongoing process rather than a one-time event, supported by periodic reporting, review mechanisms, and stakeholder consultations. Such a structured, multi-layered approach would ensure that the Dialogue is not merely deliberative, but norm-generating, implementation-oriented, and responsive to technological evolution.

Global AI governance discussions continue to reflect an imbalance in participation, with several critical voices underrepresented. First, developing countries and the Global South remain insufficiently represented in norm-setting processes. Their regulatory priorities—such as access to justice, digital infrastructure gaps, and affordability—often differ from those of technologically advanced jurisdictions. Inclusion can be strengthened through funded participation, regional consultations, and capacity-building initiatives. Secondly, legal system actors at the grassroots level—including trial court judges, local bar practitioners, and mediators—are rarely part of these discussions. Their practical insights on how AI interacts with real dispute resolution processes are essential. Structured consultation forums and pilot programmes can help integrate their perspectives. Thirdly, small and early-stage innovators, including legal-tech startups, are often overshadowed by large technology companies. Yet, they are key drivers of accessible and cost-effective solutions. Dedicated innovation tracks, startup roundtables, and regulatory sandboxes can ensure their participation. Fourthly, civil society groups representing vulnerable and marginalised communities—including those affected by algorithmic bias—require stronger representation. Their inclusion is vital to ensure that governance frameworks address real-world harms and inequities. Finally, there is limited representation from interdisciplinary experts, particularly those at the intersection of law, technology, and ethics. AI governance requires integrated perspectives rather than siloed approaches. To address these gaps, the Dialogue should adopt inclusive design mechanisms—such as multilingual engagement, hybrid participation formats, and targeted outreach—ensuring that AI governance evolves as a genuinely global and representative process.

Meaningful engagement in the AI Dialogue requires formats that move beyond static statements toward interactive, problem-oriented participation. First, scenario-based simulations can be highly effective. Participants could engage with real-world case studies—such as AI-assisted dispute resolution or cross-border data conflicts—to test how proposed principles operate in practice. This grounds discussions in legal and operational realities. Secondly, multi-stakeholder roundtables with defined outputs should replace purely open-ended discussions. Each group (States, industry, academia, civil society) can be tasked with producing short, actionable recommendations on specific themes such as liability or transparency. Thirdly, the Dialogue could incorporate "policy labs" or co-drafting sessions, where participants collaboratively develop model clauses, guidelines, or frameworks. This ensures that outcomes are not merely declaratory but implementation-oriented. Fourthly, regulatory sandbox showcases would allow startups and institutions to demonstrate working AI systems within controlled legal parameters. This bridges the gap between theory and practice and encourages innovation within compliant boundaries. Fifthly, the use of digital platforms for continuous engagement—including pre-Dialogue consultations and post-session feedback loops—can ensure broader participation, especially from stakeholders unable to attend physically. Finally, cross-disciplinary panels combining legal experts, technologists, and policymakers can foster integrated thinking, avoiding siloed approaches to governance. In essence, the Dialogue should be structured as a participatory and solution-driven process, where stakeholders not only deliberate but actively co-create practical governance tools.

Several existing policies and practices offer concrete, implementable models for effective AI governance. First, the European Union's AI Act represents a risk-based regulatory approach, classifying AI systems by levels of risk and imposing proportionate obligations. Its emphasis on high-risk systems, documentation, and human oversight provides a structured compliance model that can be adapted across jurisdictions. Secondly, UNESCO's Recommendation on the Ethics of AI (2021) establishes a globally endorsed normative framework grounded in human rights, transparency, and accountability. It is particularly valuable for States seeking soft-law guidance aligned with international legal principles. Thirdly, the OECD AI Principles promote trustworthy AI through fairness, robustness, and explainability, while encouraging international cooperation. These principles have influenced multiple national strategies and provide a common baseline for policy alignment. From a practical standpoint, regulatory sandboxes-as adopted in jurisdictions like the UK and Singapore-offer controlled environments where AI systems can be tested under regulatory supervision. This approach balances innovation with compliance and is particularly useful for emerging sectors such as legal-tech. Further, algorithmic impact assessments (AIAs), used in countries like Canada, provide a structured method to evaluate risks before deployment. They enhance accountability by requiring documentation, risk scoring, and mitigation strategies. Finally, sector-specific platforms, including online dispute resolution (ODR) systems, demonstrate how AI can be integrated responsibly to improve access to justice, provided safeguards such as transparency and human review are maintained. Collectively, these approaches illustrate that effective AI governance requires a combination of principle-based frameworks, risk-based regulation, and practical implementation tools, ensuring both legal compliance and innovation.