Dr Lamprecht

A major success would be to foster a discussion that offers a novel, precise perspective on the frictions between AI development and human rights-based governance. To achieve this, the Dialogue could especially dismantle existing narratives and enable a fundamental reality check. 1. Reversing the "Lagging Law" Narrative The discussion should move past the claim that law lags behind AI acceleration. I propose reversing the narrative and asking whether AI development is lagging behind the law. Taking copyright as an example: laws exist, but tech companies often disregard them, justifying their actions by promising long-term utility. This stance is a clear disrespect for existing social contracts. Our legal systems are the agreed models of our world; they define the boundaries of those contracts. Existing court cases are evidence that people perceive AI development as a breach of the contract. Success would be a discussion focused on actively leveraging existing legal frameworks. While the topic of regulation is highly problematic and change might be too slow, we could accelerate rebalancing our social-technical reality through a fundamental legal reality-check mechanism. It would accelerate the understanding that technical governance does not inadvertently bypass existing judicial oversight or the separation of powers. 2. The Narrative of AI's Capabilities Success means asking a critical question: Does current generative AI development accelerate progress by using society as its sandbox, or does it accelerate humanity toward the SDGs? An example is the shifting industry narrative regarding the nature of hallucinations. When a leading tech company suddenly admits that hallucinations are an inherent capability rather than a failure, it signals a shift in the metrics for AI governance and accountability. Success would be a reality check of the boundaries that AI narratives must respect to fulfil their promises.

• Interoperability of governance approaches
• Transparency, accountability, and human oversight
• Open-source software, open data and open AI models
• Protection and promotion of human rights

My selection focuses on the structural legibility of AI governance. Open-source and Open Models (4g) I prioritise Open-source and Open Models (4g) because legal systems are, by definition, the original open-source models of human societies. The Rule of Law requires that the rules governing society be public, auditable, and contestable. As AI becomes the functional "law" of our digital reality, closed-source models represent a form of secret justice. I select 4g to ensure that technical governance remains as transparent as the legal systems it seeks to mirror. Transparency and Oversight (4f) This links directly to Transparency and Oversight (4f). You cannot have robust human Oversight or judicial agency over a "black box." I prioritise these together to prevent technical metrics from bypassing the separation of powers. Oversight is a structural legal requirement, not a technical disclosure. Interoperability (4d) Interoperability must mean more than technical data flows. It must include legal compatibility with existing frameworks - ensuring that AI governance decisions do not operate outside established judicial and regulatory structures. Protection and Promotion of Human Rights (4e) Human rights law already applies to AI development; however, there is an implementation gap. Selecting this theme affirms that the Dialogue should treat human rights obligations as enforceable commitments through existing laws rather than aspirational principles. Together, these four priorities make the case that the Dialogue should focus on enforcing existing social contracts, not only on creating new frameworks.

Missing Issue: Legal Coherence - Enforcement of Existing Obligations The four themes collectively address what AI governance should look like. What is absent is a more immediate question: whether AI development already complies with existing legal obligations including human rights law, copyright, and the separation of powers. This is an enforcement gap, as regulations exist. The question is whether they are being honoured - and whether the Dialogue treats their enforcement as a governance tool in its own right. This gap has structural consequences. If thematic outcomes on interoperability, open source, or technical standards are finalised without reference to existing legal frameworks, the Dialogue risks producing recommendations that are internally coherent but incompatible with established social contracts. A practical response would be to introduce a cross-cutting reflection moment in the Way Forward session - a structured prompt asking whether thematic recommendations are compatible with existing legal obligations. This requires no new institution, but simply creates a space for Member States to reflect.

Legal sector In the legal and governance sector, we keep repeating the Collingridge Dilemma — the observation, made in 1980, that by the time the societal impacts of a technology are fully understood, it is often too entrenched to steer. It creates a systemic gap where technical metrics quietly replace our social contracts, allowing industry narratives to treat the social contract as a secondary consideration rather than the primary blueprint. This is particularly evident in the deployment of generative AI applications that treat society as a sandbox for unverified experiments. The opportunity to overcome the dilemma is to apply a structural compatibility test at the design phase — demanding that technical protocols respect existing legal constraints before deployment, rather than after entrenchment. This is what a forward-looking Dialogue could achieve: positioning law not as a reactive passenger, but as the primary architecture of our digital future. Synchronising technical advances with our social contract ensures the rule of law remains intact — regardless of the medium.

The Dialogue is uniquely positioned to introduce a disciplined question into international cooperation on AI governance: are the technical recommendations we are producing compatible with the legal obligations we have already agreed upon? By opening space for this question in the Way Forward session, before conclusions are finalised, the Dialogue can shift international cooperation from a search for new frameworks to a coherent application of existing ones. This is a procedural change, not an institutional one. It requires no new bodies. It simply ensures that technical progress and the rule of law move in the same direction.

The Dialogue should build on the foundational mandates of the Global Digital Compact and the Pact for the Future, while connecting to the UN 2.0 Quintet of Change as an operational framework. The Quintet's pillars of data, digital, innovation, foresight, and behavioural science offer practical tools to move governance from reactive regulation to deliberate design. The UN Virtual Worlds and Citiverse initiatives offer a concrete testing ground for this approach. These initiatives represent the transition from abstract policy to real-world implementation, where emerging technical standards for digital identity, property, and interaction are actively being shaped. The added value the Dialogue brings is an open invitation to reflect on these interconnected initiatives against existing legal obligations before drawing conclusions, building on the expertise already present in the room. This is what could distinguish the Dialogue from the many forums already addressing AI governance: a deliberate effort to ensure that international cooperation produces recommendations compatible with the social contracts we have already agreed upon.

The multi-stakeholder format is the Dialogue's greatest asset, as it brings together the legal, technical, and policy expertise needed to discuss whether AI governance recommendations are coherent with existing frameworks. The proposed structure is a solid foundation. One addition could increase its impact: the Way Forward should include an open invitation for Member States and stakeholders with relevant legal expertise to pilot a compatibility reflection, asking whether thematic outcomes are coherent with existing legal frameworks. This requires no new mechanism and no cross-jurisdictional assessment. It simply creates space for those with the expertise and willingness to demonstrate what legal coherence looks like in practice, building a body of examples that future iterations of the Dialogue can build upon.

The most underrepresented perspectives are legal practitioners and legal tech experts. While policy is discussed broadly, the experts who actually practise the law and those who translate legal principles into technical code are currently missing from the design phase. Their inclusion could be achieved through dedicated expert panels within the thematic breakouts and by connecting the Dialogue to legal communities. This would ensure that the Dialogue draws on the expertise needed to reflect technical outcomes against existing legal obligations and that this expertise is carried forward into future iterations.

To foster dynamic engagement, I recommend utilising the AI for Good Neural Network app. This platform allows participants to connect directly and maintain a continuous overview of the Dialogue. Crucially, it demonstrates a converging effort within the UN system, linking the Dialogue to existing initiatives instead of creating siloed efforts.

Many legal tech initiatives illustrate what effective AI governance looks like when legal practitioners are at the centre of the process. For example, a pioneer in Switzerland is The Weblaw Forum Legal Tech that brings together judges, lawyers, in-house counsel, and legal tech experts to examine AI deployment in legal practice. In Austria, the Future-Law Legal Tech Konferenz operates similarly, convening practitioners, the Austrian Ministry of Justice, and legal tech providers to address AI adoption from the perspective of those who apply the law daily. There are two examples among many across different countries and jurisdictions. It would be my honour to bring these communities to the table.