With AI We Innovate Medicine Association, Shaqra University

To be considered a success, the first Global Dialogue on AI Governance must move beyond abstract ethical alignment toward a model that preserves human agency through what may be described as cognitive governance. Success would be defined by three pivotal outcomes: 1. Recognition of Cognitive Sovereignty The Dialogue should establish that AI responsibility extends beyond algorithmic fairness to include interpretive sovereignty—the fundamental human right to understand, question, and meaningfully engage with AI-driven decisions. Transitioning from opaque "black-box" systems toward designs that prioritize human comprehension is essential to prevent cognitive alienation and ensure accountable use. 2. Standardizing the "Meaning Gap" as a Systemic Risk A successful outcome would formally recognize the "meaning gap"—the disconnect between system outputs and human understanding—as a governance risk. A decision that is technically accurate but not humanly interpretable should be treated as a limitation in safety, oversight, and trust. Addressing this gap is critical for sustainable adoption. 3. Institutionalizing the Human-in-the-Logic The Dialogue should produce a roadmap ensuring that humans remain active participants within decision processes, not passive executors. This includes: redefining leadership as the responsibility to ensure shared understanding of AI-assisted decisions; expanding evaluation frameworks beyond quantitative metrics to include the quality of human comprehension and engagement. Ultimately, success will be measured by whether global governance frameworks preserve the human capacity to understand, interpret, and take responsibility for decisions. By doing so, AI can evolve as a tool for cognitive augmentation—rather than a substitute for independent human judgment.

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

My selection of these four thematic areas is intended to ground the proposed "Ten Principles of Human-Centered Governance" within the United Nations framework, ensuring that AI remains a tool for human empowerment rather than a substitute for human judgment. Transparency, accountability, and human oversight represent the cornerstone of this approach. Effective accountability requires what may be described as interpretive sovereignty-the human ability to understand, question, and meaningfully engage with AI-driven decisions. Without this, oversight risks becoming procedural rather than substantive. The social, ethical, and cultural implications of AI provide a critical space to address the emerging "meaning gap" as a systemic risk-where decisions may be technically accurate but not sufficiently understood. This dimension ensures that AI adoption remains aligned with cultural contexts and supports inclusive, context-aware governance. Protection and promotion of human rights extends this perspective by framing cognitive integrity as a foundational concern. Preserving the human capacity to interpret, evaluate, and take responsibility for decisions is essential to maintaining dignity and agency in increasingly automated environments. Finally, safe, secure and trustworthy AI must be understood beyond technical robustness. Trust is not achieved through performance alone, but through alignment with human understanding. In this sense, safety and trust are strengthened when systems are designed to support cognitive clarity and balance between human judgment and machine outputs. Together, these priorities advance a human-centered model of AI governance that safeguards understanding, reinforces accountability, and ensures that technological progress does not come at the expense of human agency.

Yes. A critical emerging issue not fully captured across the listed themes is the need to govern the human-AI cognitive interface, particularly through what may be described as cognitive sovereignty and the "meaning gap" as a systemic risk. While current frameworks address technical performance, ethics, and accountability, they often assume that AI outputs-if accurate and transparent-are also meaningfully understood. In practice, this is not always the case. First, cognitive integrity should be recognized as a foundational consideration. AI governance must ensure that individuals can meaningfully align their reasoning with the decisions they adopt, preserving human agency and responsibility. Second, the "meaning gap"-where system outputs are technically accurate but not sufficiently interpretable-should be treated as a systemic governance concern. Such gaps can weaken oversight, diffuse accountability, and undermine trust, even in otherwise compliant systems. Third, there is an emerging risk of cognitive alienation, where excessive reliance on AI may reduce independent human judgment and diminish active engagement in decision-making processes. Fourth, governance should increasingly be understood as cognitive design. This implies that interpretability and human comprehensibility are embedded at the system design stage, rather than addressed only after deployment. Finally, the evolving role of leadership in AI-enabled environments includes ensuring shared understanding of AI-supported decisions, not only their implementation. Addressing these cross-cutting issues would strengthen existing themes by ensuring that AI governance not only regulates systems, but also preserves the human capacity to understand, interpret, and take responsibility for decisions.

In the context of the academic and healthcare sectors (particularly pharmacy) in Saudi Arabia, current AI governance gaps present both significant challenges and strategic opportunities. Significant Challenges A primary concern is the "meaning gap" as a systemic risk. In pharmaceutical research and clinical decision-making, AI systems may produce technically accurate outputs that are not fully interpretable by researchers or practitioners. This creates a disconnect between adoption and understanding, potentially affecting the quality and accountability of scientific decisions. There is also an emerging risk of cognitive alienation, where increasing reliance on AI systems may gradually reduce critical thinking among researchers and students. Over time, this could weaken intellectual independence and diminish active engagement in knowledge generation. Additionally, the erosion of cognitive balance—between data-driven outputs and human expertise—may lead to the decline of traditional analytical and evaluative skills within academic environments. Strategic Opportunities At the same time, these gaps present a unique opportunity to advance more human-centered governance approaches. Saudi Arabia's transformation agenda provides a foundation to redefine leadership as ensuring collective understanding of AI-supported decisions, strengthening trust and responsible adoption. There is also an opportunity to position governance as cognitive design, integrating human interpretability and comprehension into system development from the outset, rather than addressing them post-deployment. Finally, intergenerational engagement can be leveraged to support inclusive AI adoption, ensuring that differences in technological familiarity are addressed through adaptive policies and education. Addressing these challenges through human-centered governance approaches can enable the Saudi academic and healthcare sectors to contribute a globally relevant model—one that balances innovation with the preservation of human understanding, agency, and responsibility.

The Global Dialogue on AI Governance can serve as a definitive platform for advancing international cooperation by extending alignment beyond technical standards toward a cognitive governance framework. 1. Establishing a Universal "Cognitive Minimum" The Dialogue can facilitate global consensus on what may be described as interpretive sovereignty—the principle that human understanding is a non-negotiable component of responsible AI. By recognizing that decisions must be meaningfully interpretable to those who rely on them, international cooperation can ensure the protection of cognitive integrity across borders and reduce the risk of a global "meaning gap." 2. Harmonizing Standards through Cognitive Design Rather than fragmented regulatory approaches, the Dialogue can promote governance as cognitive design. This involves setting shared benchmarks where AI systems are evaluated not only for fairness and safety, but also for their "human-readability" and their ability to maintain a balance between algorithmic outputs and human reasoning. Such alignment would strengthen trust and interoperability across jurisdictions. 3. Integrating Cross-Cutting Cognitive Risks The Dialogue can provide a mechanism to formally recognize emerging risks—such as misunderstanding and overreliance—as governance concerns. Addressing these risks at a global level would enhance consistency in how safety, accountability, and oversight are implemented. 4. Enabling Inclusive, Multi-Stakeholder Engagement Finally, the Dialogue can ensure that diverse sectors and regions contribute to shaping governance frameworks that reflect different cultural and institutional contexts, while maintaining shared principles of human-centered AI.

The AI Dialogue should build upon existing international initiatives while addressing key gaps by integrating a more human-centered dimension of governance. Existing Mechanisms to Connect With The Dialogue can strategically connect with established frameworks such as the UNESCO Recommendation on the Ethics of AI, the OECD AI Principles, and the Global Partnership on AI (GPAI). These initiatives have advanced important foundations in ethics, safety, and trustworthy AI. The added contribution of the Dialogue would be to extend these efforts from high-level principles toward a more operational perspective that incorporates human understanding. For example, while the OECD emphasizes trust, the Dialogue can strengthen this by ensuring that trust is grounded in meaningful human interpretability, not only system performance. Similarly, collaboration with GPAI can help integrate human-centered considerations into ongoing technical and policy research. Added Value of the AI Dialogue The unique value of the Dialogue lies in shifting focus from what AI systems do to how humans understand and engage with them. This includes: recognizing the "meaning gap"—where outputs are technically valid but not sufficiently interpretable—as a governance concern relevant to safety, accountability, and trust; promoting governance approaches that embed human interpretability into system design, rather than addressing it only after deployment; strengthening the role of leadership and institutions in ensuring shared understanding of AI-supported decisions. By building on existing initiatives while introducing this human-centered, cross-cutting perspective, the AI Dialogue can enhance coherence across global frameworks and support governance models that preserve human agency, responsibility, and meaningful participation in AI-driven environments.

To ensure the Global Dialogue on AI Governance is truly inclusive, it must move beyond traditional diplomatic engagement toward a multi-layered ecosystem of stakeholders, enabling meaningful participation across sectors. 1. Stakeholder Contributions Academia and Research Institutions can provide interdisciplinary frameworks that translate conceptual approaches into actionable governance models, particularly in areas such as interpretability and human–AI interaction. Civil Society and Think Tanks can ensure that governance reflects societal values, safeguarding human agency and preventing the marginalization of human judgment in AI-driven environments. Private Sector Actors play a critical role in operationalizing governance principles by embedding human interpretability and usability into system design and deployment. Applied Sectors (e.g., healthcare, education) can contribute real-world insights on how AI systems are understood, adopted, and trusted in high-impact contexts. 2. Recommended Format and Structure Thematic and Cross-Cutting Working Groups to address key governance areas while integrating human-centered considerations across all themes. Cognitive Impact Assessments to evaluate not only technical performance, but also the extent to which AI systems are interpretable and usable by their intended users. Intergenerational Dialogue Platforms to ensure policies reflect diverse levels of technological familiarity and support inclusive adoption. Decentralized and Hybrid Engagement Models that combine global dialogue with local, context-specific consultations. Expanded Evaluation Frameworks that complement quantitative metrics with qualitative assessments of understanding, trust, and effective use.

In the current global landscape, an important underrepresented perspective is the human-centered, cognitive dimension—approaches that view AI not only as a technical system, but as a transformation affecting human agency, understanding, and identity. While technical, legal, and policy experts are well represented, several voices remain less visible: 1. Underrepresented Perspectives Interdisciplinary researchers working at the intersection of cognition, human behavior, and technology. Their work is essential to understanding how individuals interpret and interact with AI systems, particularly in relation to decision-making and responsibility. Non-technical leadership perspectives, including leaders in education, healthcare, and public institutions, who focus on collective understanding, trust, and societal adoption rather than purely technical performance. Generationally diverse users, including both younger and older populations, who experience AI systems differently and may face distinct challenges in understanding and engaging with them. 2. Pathways to Inclusion To address these gaps, the AI Dialogue could: Integrate interdisciplinary expertise—including cognitive science, behavioral science, and related fields—into policy discussions and system design considerations. Promote human-centered governance approaches, ensuring that AI systems are not only compliant and efficient, but also understandable and usable across diverse contexts. Establish inclusive dialogue platforms that actively engage different generations and user groups to capture varied experiences of AI adoption. Encourage design-stage considerations that prioritize interpretability and usability, ensuring that human understanding is embedded from the outset.

To foster meaningful and dynamic engagement, the AI Dialogue should move beyond static presentations toward interactive formats that actively examine how humans understand and engage with AI systems. 1. Scenario-Based Cognitive Simulation Labs Stakeholders participate in structured simulations using real-world AI use cases (e.g., healthcare, education), allowing them to interact with AI-driven decisions. This format helps identify gaps between system outputs and human understanding in real time, and evaluates whether decisions are meaningfully interpretable and adoptable. 2. Human-Centered Design Workshops Interdisciplinary sessions that bring together technical experts, policymakers, and applied professionals to define essential aspects of human judgment that must be preserved in AI systems. These workshops can help ensure that governance frameworks incorporate human interpretability as a design requirement. 3. Interpretability Review Panels Structured review formats where AI systems are assessed not only for accuracy and compliance, but also for their ability to explain decision logic to diverse stakeholders. This supports accountability and strengthens trust across different user groups. 4. Intergenerational Engagement Forums Dialogue platforms that bring together participants with varying levels of technological familiarity to explore how different groups perceive and understand AI systems, ensuring inclusive and context-sensitive governance. 5. Expanded Evaluation Challenges Innovation-focused sessions that encourage the development of qualitative assessment approaches, complementing traditional quantitative metrics by evaluating understanding, usability, and effective human engagement.

A notable example of an emerging policy approach is the integration of a human-centered "cognitive governance" model within national AI strategies, such as proposals linked to the Saudi Responsible AI Policy. This approach complements existing frameworks by extending governance beyond technical compliance toward ensuring meaningful human engagement with AI systems. Concrete Policy Approaches Cognitive Integrity: Evaluating decision quality not only by algorithmic accuracy, but by the extent to which users can understand and meaningfully adopt the outcome. Interpretability and Right to Review: Establishing policies that ensure individuals can access, understand, and question the logic behind AI-driven decisions, strengthening accountability. Leadership for Shared Understanding: Expanding the role of leadership to include ensuring that AI-supported decisions are clearly understood within organizations, supporting responsible use. Recognition of the "Meaning Gap": Treating the gap between system outputs and human understanding as a governance concern relevant to safety, oversight, and trust. Effective Practices and Platforms Design-Stage Integration: Embedding interpretability and usability into system architecture from the outset, rather than addressing them only after deployment. Balanced Human-AI Interaction Models: Ensuring that AI augments human expertise while preserving independent judgment. Inclusive Engagement Platforms: Incorporating diverse user perspectives, including intergenerational dialogue, to improve adoption and trust. Expanded Evaluation Frameworks: Complementing quantitative performance metrics with qualitative assessments of understanding, usability, and effective decision-making.