universty malek asaadi tanger morocco

Here's my take, — informed by the Dialogue's mandate and the geopolitical landscape heading into Geneva this July: The first Global Dialogue on AI Governance would succeed if it achieves three things: 1. Genuine inclusivity beyond symbolism. The Dialogue's premise — that every country gets a seat at the AI governance table — is powerful Center for Strategic and International Studies, but success means Global South voices actually shape outcomes, not merely attend. Concrete commitments to capacity-building, technology transfer, and representation of African, Arab, and small-state perspectives in follow-up mechanisms would distinguish this from past multilateral rituals. 2. Interoperability over harmonization. Rather than pursuing a single global framework (politically impossible given U.S. opposition to multilateral AI governance Center for Strategic and International Studies and China's competing vision), success would look like a practical roadmap for making diverse national and regional approaches — the EU AI Act, China's tiered regulations, emerging African frameworks — legible to one another. Mutual recognition mechanisms, shared taxonomies of risk, and agreed minimum transparency standards would be realistic and valuable deliverables. 3. Linking governance to the Scientific Panel. The Independent International Scientific Panel on AI should present its first annual assessment during the Dialogue Simoninstitute. A successful outcome would anchor political discussions in that evidence base — particularly on frontier AI risks and the uneven distribution of AI's benefits — rather than letting the conversation remain at the level of aspirational principles. The international community must move beyond principles and turn evidence into commitments, and commitments into cooperation. CyberPeace Institute A less visible but equally important marker of success: establishing clear pathways for non-state stakeholders — civil society, researchers, the private sector — to contribute substantively to the second Dialogue in 2027, avoiding the trap of state-only deliberation on a technology that transcends borders by design. If Geneva produces even a modest but actionable framework for interoperability plus meaningful Global South agency, it will have justified its existence.

• Safe, secure and trustworthy AI
• AI capacity-building
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Interoperability of governance approaches

I would prioritize Interoperability of governance approaches, for one fundamental reason: it is the enabling condition for the other three. Without interoperability, "safe, secure and trustworthy AI" remains fragmented - each jurisdiction defining safety by its own standards, creating regulatory arbitrage opportunities and compliance chaos for cross-border AI systems. Capacity-building efforts lose coherence if the governance models being transferred to developing nations are mutually incompatible. And meaningful engagement with AI's social, ethical, and cultural implications requires a shared vocabulary - otherwise these conversations happen in parallel silos that never converge. The current landscape makes this urgent. The EU AI Act imposes risk-based obligations with extraterritorial reach. China operates a sector-specific regulatory stack. The US has retreated from multilateral commitments. African states are drafting national strategies with limited coordination. The result is a governance patchwork where a Moroccan FinTech deploying AI-based credit scoring could simultaneously face EU adequacy requirements, African Union data sovereignty principles, and domestic BAM sandbox rules - with no mechanism for reconciling them. Interoperability does not mean harmonization. It means building translation layers: shared risk taxonomies, mutual recognition protocols, agreed transparency baselines. This is where the PNCS (Pluralisme Normatif Contextuel Stratégique) approach becomes relevant - accepting normative pluralism as a feature rather than a bug, while creating strategic bridges between regulatory traditions. This also has direct equity implications. When governance frameworks are interoperable, Global South countries can participate in the AI economy without being forced into regulatory mimicry of European or American models. They can develop contextually appropriate approaches that remain legible - and therefore credible - to international partners. Geneva should produce not a single rulebook, but a Rosetta Stone - the mechanisms through which diverse approaches can communicate, coexist, and progressively converge where convergence serves humanity.

Yes. Two critical cross-cutting issues are conspicuously absent: 1. AI and Normative Sovereignty in the Global South The listed themes implicitly assume a universal consensus on whose human rights framework, whose transparency standards, and whose open-source ecosystem defines the baseline. They do not address the power asymmetry embedded in AI governance itself. When the EU AI Act's extraterritorial provisions effectively export European normative choices to African and Arab markets - without reciprocal influence - the governance process reproduces the dependency it claims to remedy. What is missing is an explicit theme on regulatory sovereignty and contextual adaptation: the right of developing nations to craft AI governance reflecting their own legal traditions, cultural values, and development priorities - not as exceptions to a Northern default, but as legitimate normative contributions. Morocco's position between European, African, and Arab regulatory traditions illustrates this perfectly. The country should not be reduced to a compliance-taker; it can function as a normative mediator, translating between governance philosophies rather than simply importing one. 2. AI, Linguistic Justice, and Epistemic Inclusion The themes mention "linguistic implications" elsewhere but do not address the foundational issue: current AI systems encode a hierarchy of languages that determines who benefits and who is excluded. Arabic, Amazigh, Darija, Wolof, Swahili - these are not edge cases. They represent billions of people whose knowledge systems, legal concepts, and cultural expressions are systematically underrepresented in training data, evaluation benchmarks, and governance discussions alike. This is not merely a technical gap. It is an epistemic one. AI systems that cannot reason adequately in Arabic legal terminology or capture African oral knowledge traditions will produce governance recommendations blind to the realities of the majority world. Linguistic justice must be treated as infrastructure, not afterthought - a precondition for every other theme on this list to deliver equitable outcomes.

Morocco stands at a unique crossroads — simultaneously exposed to governance gaps and positioned to bridge them. The interoperability challenge is immediate and concrete. Moroccan financial institutions deploying AI-driven KYC/AML and credit scoring systems must navigate Bank Al-Maghrib's sandbox framework, CNDP data protection rules modeled on GDPR, emerging African Union data sovereignty principles, and EU AI Act extraterritorial obligations when serving European counterparts. No reconciliation mechanism exists between these regimes. The result: compliance costs disproportionately burden Moroccan FinTechs compared to European competitors operating under a single framework, directly undermining the financial inclusion objectives of the SNIF 2023-2026 strategy. The normative sovereignty gap is structural. Morocco's Digital Morocco 2030 strategy and Maroc IA 2030 ambitions are being developed while the EU AI Act effectively pre-determines regulatory choices for any Moroccan company touching European markets. This creates a two-tier system — innovation shaped by Brussels for export-facing sectors, and a domestic vacuum for everything else. The country risks becoming a regulation-taker rather than a regulation-maker, despite possessing the institutional maturity and strategic positioning to contribute original governance models. Yet the opportunities are significant. Morocco's trilingual legal infrastructure, its geographic and institutional position bridging Africa, the Arab world, and Europe, and initiatives like the JAZARI institutes and D4SD Hub create genuine potential for normative mediation. The country could pilot interoperability frameworks that translate between EU, African, and Arab governance approaches — a living laboratory for the pluralist model the Global Dialogue needs. The linguistic dimension compounds everything. AI systems performing inadequately in Darija and Amazigh exclude rural populations from FinTech services, e-government platforms, and digital justice — precisely the populations financial inclusion policies target. Until multilingual AI infrastructure is treated as governance infrastructure, Morocco's digital transformation will deepen the divides it promises to close.

The Dialogue's greatest potential lies in institutionalizing pluralism — making diverse governance approaches an asset rather than an obstacle to cooperation. First, establishing interoperability mechanisms. The Dialogue should create working groups tasked with developing shared risk taxonomies, mutual recognition protocols, and cross-jurisdictional transparency standards. Drawing from my experience deploying AI systems across financial institutions — where fraud detection and biometric models must simultaneously satisfy multiple regulatory regimes — I can attest that interoperability is not abstract theory. It is an operational necessity that practitioners solve daily without institutional support. The Dialogue can formalize what the field already demands. Second, legitimizing polycentric governance. My doctoral research develops the PNCS framework (Pluralisme Normatif Contextuel Stratégique), which argues that effective AI governance requires accepting normative plurality rather than pursuing premature convergence. The Dialogue is uniquely positioned to operationalize this — creating structured space where African, Arab, European, and Asian regulatory philosophies engage as equals. Morocco's experience navigating between EU adequacy requirements, African Union digital sovereignty principles, and domestic innovation objectives provides a concrete model for this mediation role. Third, bridging the technical-legal divide. AI governance suffers from a persistent disconnect between those who build systems and those who regulate them. My dual positioning — AI .big data and cybersecurity credentials combined with doctoral research in international business law — reflects exactly the interdisciplinary profile this Dialogue needs to cultivate. Cooperation falters when lawyers draft obligations engineers cannot implement, and when engineers deploy systems lawyers cannot evaluate. The Dialogue should institutionalize technical-legal translation capacity. Fourth, anchoring cooperation in evidence. The Scientific Panel's assessments should directly inform Dialogue deliberations, ensuring that international cooperation addresses documented risks and measurable inequities — particularly regarding linguistic exclusion, capacity asymmetries, and the concentration of AI development in a handful of nations.

The Dialogue must not duplicate — it must connect, translate, and fill gaps between existing mechanisms. Existing initiatives to build upon. The OECD AI Principles provide a valuable normative baseline, but remain anchored in developed-economy perspectives. The G7 Hiroshima Process advanced frontier AI governance yet excluded the Global South entirely. The Council of Europe AI Treaty offers the first binding international instrument but faces ratification uncertainty. The African Union's Continental AI Strategy provides a pan-African vision still lacking implementation infrastructure. Regional efforts — EU AI Act, China's tiered regulations, Singapore's FEAT principles, Morocco's BAM sandbox — represent rich regulatory experimentation that remains siloed. The Dialogue's unique added value is threefold. First, institutional bridging. No existing mechanism connects the OECD's technical standards work with the AU's development-focused approach or the Arab League's emerging digital strategies. From my experience operating across Moroccan, African, Arab, and European regulatory contexts in financial technology — deploying production ML systems for fraud detection, biometric authentication, and KYC/AML compliance — I understand firsthand that these frameworks interact at the operational level but never at the institutional level. The Dialogue can formalize these connections. Second, cybersecurity-governance integration. Current AI governance initiatives treat cybersecurity as peripheral. My expertise confirms this is dangerous — AI systems processing big data across jurisdictions create attack surfaces that no single national framework addresses. The Dialogue should connect AI governance with existing cybersecurity cooperation mechanisms, including the Budapest Convention and FIRST network. Third, legal interoperability architecture. My doctoral research in international business law identifies a critical gap: no mechanism exists for mutual recognition of AI regulatory decisions across jurisdictions. The Dialogue could pioneer frameworks analogous to international investment law's bilateral treaty architecture — creating structured bilateral and multilateral AI governance agreements that respect normative sovereignty while enabling cross-border cooperation.

The Dialogue must be structured to produce actionable outcomes, not declaratory statements. Multi-stakeholder contribution architecture. States should lead on normative commitments, but the Dialogue needs structured roles for practitioners who operate at the intersection of regulation and implementation. Technical experts with production experience — deploying ML systems for fraud detection, biometric verification, and big data analytics across regulated industries — bring irreplaceable knowledge about what governance frameworks actually demand at the operational level. My two decades deploying these systems across financial institutions demonstrate that governance designed without practitioner input produces elegant rules that collapse upon implementation. Recommended format: three-layer deliberation. Layer one — Technical Working Groups. Permanent, intersessional groups organized around concrete deliverables: shared risk taxonomies, interoperability protocols, cybersecurity baseline standards for AI systems. These should integrate dual-competency experts — professionals combining technical credentials with legal expertise in international business law — who can translate between engineering constraints and regulatory objectives. This technical-legal translation capacity is precisely what current mechanisms lack. Layer two — Regional Caucuses. Structured preparation sessions where African, Arab, Asian, European, and Latin American stakeholders develop coordinated positions before plenary deliberations. Morocco could anchor an Africa-Arab caucus reflecting its bridging position between governance traditions — operationalizing the pluralist approach my PNCS framework advocates. Layer three — Annual Plenary with binding work programme. Not another summit producing aspirational communiqués, but a structured session that adopts specific interoperability deliverables, mandates the Scientific Panel to investigate identified gaps, and assigns intersessional tasks with deadlines. Critical structural recommendation: establish a permanent Practitioners Advisory Council — industry experts, civil society technologists, academic researchers from the Global South — ensuring the Dialogue remains grounded in implementation reality rather than drifting into diplomatic abstraction.

The absence is systematic — not accidental — and reproduces the very inequities AI governance claims to address. African and Arab practitioners. Global AI governance is dominated by Euro-Atlantic voices debating frameworks designed for their contexts. African engineers building AI solutions for financial inclusion, Arab researchers developing Arabic NLP systems, Moroccan cybersecurity professionals securing cross-border financial infrastructure — these practitioners possess experiential knowledge that no Northern think tank can substitute. My own trajectory illustrates this gap: over twenty years deploying production ML systems across financial institutions in African and European contexts, yet the forums shaping AI regulation rarely include practitioners with this dual operational perspective. Linguistic communities. Darija, Amazigh, Wolof, Hausa, Amharic speakers represent hundreds of millions whose languages remain computationally marginalized. Their exclusion from AI governance is not merely representational — it is epistemic. Governance frameworks designed in English and French cannot adequately address AI systems that systematically fail communities operating in these languages. Inclusion requires funding multilingual participation infrastructure, not simply translating final documents. Small and medium enterprises from developing economies. Current consultations privilege Big Tech and well-funded civil society organizations. Moroccan FinTechs navigating BAM sandbox requirements, African startups deploying mobile credit scoring, Arab e-commerce platforms integrating AI — these actors bear governance costs disproportionately while having minimal influence over rule-setting. Interdisciplinary professionals. AI governance remains siloed between technical communities and legal communities that rarely interact substantively. Professionals bridging both — combining cybersecurity expertise with international business law research, as my own CISSP/CISM/CEH credentials alongside doctoral work demonstrate — are essential for producing implementable governance. Inclusion mechanisms: funded fellowship programmes for Global South participants, mandatory regional representation quotas in working groups, intersessional virtual consultations accommodating time zones and connectivity constraints, and dedicated Arabic and African language working sessions — not as parallel events but as integral deliberation spaces.

The Dialogue must abandon the conference-panel model that produces passive audiences and predictable statements. Regulatory Simulation Labs. Instead of abstract debates about interoperability, participants should engage in live exercises where a single AI system — a credit scoring model, a biometric authentication platform — is simultaneously evaluated against EU AI Act requirements, African Union data sovereignty principles, and Arab regulatory frameworks. My experience deploying fraud detection and KYC/AML systems across multiple jurisdictions confirms that governance contradictions become visible only through operational simulation. These labs would produce concrete interoperability gap analyses within hours rather than years of diplomatic negotiation. Reverse Mentorship Sessions. Global South practitioners brief Northern policymakers on governance realities — not the reverse. A Moroccan FinTech founder explaining BAM sandbox navigation challenges, a Kenyan mobile money engineer describing algorithmic fairness in cash-scarce economies, an Arab NLP researcher demonstrating bias in Arabic language models. This inverts the conventional knowledge hierarchy and surfaces implementation intelligence that traditional formats suppress. Technical-Legal Translation Workshops. Structured sessions pairing cybersecurity engineers with international law scholars to co-draft governance provisions in real time. Drawing from my dual positioning — technical credentials alongside doctoral research in international business law — this format addresses the fundamental disconnect producing unimplementable regulation. Participants would collaboratively stress-test proposed AI governance language against technical feasibility. Living Document Sprints. Rather than producing static outcome documents, the Dialogue should host collaborative drafting sessions where interoperability frameworks, shared risk taxonomies, and mutual recognition protocols are developed iteratively across regional caucuses — updated in real time and published as evolving working documents between annual sessions. Multilingual Deliberation Circles. Small-group discussions conducted simultaneously in Arabic, French, English, Swahili, and Mandarin — with AI-assisted interpretation — ensuring linguistic diversity shapes substance rather than merely accessing it through translation of predetermined conclusions.

Effective AI governance already exists in fragments - the challenge is connecting and scaling what works. Regulatory sandboxes as governance laboratories. Bank Al-Maghrib's 2021 FinTech sandbox represents a model worth replicating: controlled environments where AI-driven financial services - credit scoring, mobile payments, KYC/AML automation - are tested against regulatory requirements before full deployment. From my experience deploying production ML systems across financial institutions, sandboxes solve a critical problem: they allow regulators to understand what they are regulating before regulating it. Singapore's FEAT principles and the UK's FCA sandbox offer complementary approaches that the Dialogue should map for interoperability. Risk-based tiered classification. The EU AI Act's four-tier risk framework provides useful architecture despite its extraterritorial overreach. My doctoral research identifies both its strengths - proportionality, legal certainty - and its blind spots: the classification criteria reflect European risk perceptions that may not translate to African or Arab contexts where financial exclusion poses greater societal risk than surveillance concerns prioritized by European legislators. Cross-jurisdictional cybersecurity frameworks. The NIST AI Risk Management Framework and ISO/IEC 42001 offer concrete governance tools for AI system security. My practice confirms these standards work operationally - but only when adapted contextually. A Moroccan financial institution securing biometric authentication systems faces threat landscapes distinct from European counterparts. Governance frameworks must accommodate this variance. Open-source governance tools. OECD's Catalogue of AI Tools and Metrics democratizes governance capacity. Morocco's JAZARI institutes and D4SD Hub could serve as distribution nodes, adapting these tools for African and Arab regulatory contexts. The missing practice: normative mediation platforms. No existing mechanism facilitates structured translation between governance philosophies. My PNCS framework proposes precisely this - strategic pluralism operationalized through institutional mechanisms where jurisdictions negotiate interoperability bilaterally rather than surrendering sovereignty multilaterally. The Dialogue should pilot this approach.