Creative Humanity AI Education

Success would mean three concrete outcomes: First, formal recognition of education hubs as governance infrastructure alongside regulation and technical standards. The gap between international principles (UNESCO, OECD) and national enforcement (EU AI Act) cannot be bridged by more frameworks alone – it requires translational capacity that turns principles into practitioner competence. Education hubs provide this missing layer. Second, Member State commitment to develop rights-functional AI taxonomies that classify systems by their impacts on privacy, non-discrimination, due process, and remedy – not by technical architecture. Current technique-based definitions enable "definitional arbitrage" where regulated actors claim their systems are "statistical analysis" rather than AI, escaping accountability. What matters for rights protection is whether a system affects a person's rights, not whether it uses transformers or gradient boosting. Third, acknowledgment that AI governance shapes not only systems but also the humans who design, deploy, and interact with them. This cultural dimension – what we call "bidirectional training" – is missing from current governance discourse. How practitioners speak to AI, how organizations implement frameworks, how communities participate in governance: these human practices are the substrate on which any durable regime rests. Success is not another declaration of principles. It is actionable infrastructure: funded hubs translating frameworks into capacity, clear definitions enabling consistent rights protection, and recognition that governance is cultural work, not only regulatory work.

• Protection and promotion of human rights
• Transparency, accountability, and human oversight
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• AI capacity-building

These four priorities form an integrated whole. Protection and promotion of human rights and transparency, accountability, and human oversight are inseparable from the definitional challenge we address in our submission. Without legal clarity on what constitutes AI, rights-based protections become unenforceable - companies exploit definitional ambiguity to escape accountability. Rights-functional taxonomies, which classify systems by their impacts rather than technical methods, would close this gap. Social, economic, ethical, cultural implications matter because AI governance is not only regulatory - it is cultural. How practitioners interact with AI systems shapes institutional norms, professional ethics, and ultimately the societies these systems serve. Our newsletter research on "bidirectional training" demonstrates that every human-AI interaction flows both ways: those who treat AI as mere optimization receive efficiency; those who bring authentic human moments contribute something that endures. This cultural substrate determines whether governance regimes remain abstract principles or become lived practice. AI capacity-building is the mechanism that connects principles to practice. Education hubs provide the translational layer between international frameworks (UNESCO, OECD, Council of Europe) and sector-specific implementation. Practitioners in our workshops consistently ask: "How do I apply these principles tomorrow at work?" Without structured capacity-building infrastructure - funded, sustained, replicable - this gap persists. Together, these priorities address the governance question from definition to implementation: clear rights-based definitions, cultural recognition of governance as human practice, and institutional capacity to translate principles into competence.

Yes - the institutional gap between soft law and hard law. Current governance discourse oscillates between universal normative frameworks (UNESCO Recommendation, OECD Principles) that lack enforcement mechanisms, and binding instruments (EU AI Act, national regulations) that fragment across jurisdictions. Neither layer alone can succeed. Universal frameworks remain aspirational without implementation pathways; national regulations create compliance costs without shared learning. The missing layer is intermediate governance infrastructure: institutions that translate international principles into situated, sector-specific capacity while enabling cross-jurisdictional knowledge exchange. Education hubs occupy this layer. They are neither regulators nor standard-setters - they are translators, bridging normative intent and organizational practice. This institutional gap manifests practically: a hospital deploying diagnostic AI knows it must respect patient rights, but cannot navigate the distance between a UNESCO principle and a procurement decision. A startup building content moderation tools recognizes the need for transparency but lacks frameworks for implementing it across different legal regimes. Education hubs exist to close these gaps through practitioner training, multi-stakeholder dialogue, and open-source methodologies that enable replication without centralization. Recognizing this intermediate layer would shift governance from a binary (principles vs. enforcement) to a triad (principles, capacity-building infrastructure, enforcement). Member States could fund hubs as governance investments, not merely development projects. Multi-stakeholder participation would occur not only in high-level consultations but in sustained, localized implementation work. The emerging issue is institutional architecture: what governance layers exist between international consensus and national enforcement, and how do we fund, scale, and coordinate them?

In European education and cultural production sectors, governance gaps manifest as practitioner paralysis. The challenge: educators, filmmakers, and cultural workers recognize AI's transformative potential but lack operational frameworks to implement ethical principles. They attend workshops asking "How do I apply UNESCO guidelines in my classroom tomorrow?" or "Which GDPR provisions apply to this student-facing chatbot?" The gap between international frameworks and daily practice leaves them choosing between inaction (rejecting AI entirely) or improvisation (deploying systems without rights-based safeguards). This paralysis has regional consequences. Europe positions itself as the rights-based governance leader, yet practitioners lack the institutional support to translate that leadership into lived practice. The EU AI Act provides regulatory clarity for high-risk systems but offers limited guidance for the hundreds of moderate-risk applications educators and cultural producers encounter daily. Meanwhile, US platforms offer frictionless adoption with minimal rights protections, creating competitive pressure to abandon European values for operational convenience. The opportunity: Europe's strong civil society tradition, multilingual capacity, and cross-border networks position it uniquely to build intermediate governance infrastructure. Mediterranean regions connecting Europe, North Africa, and the Middle East can serve as bridges for governance dialogue beyond the US-China-EU trilateral framework that currently dominates discourse. Concretely, we see demand: our workshops consistently oversubscribe, our newsletter maintains 44-50% open rates in a sector where 20% is typical, and practitioners request ongoing training beyond single sessions. The gap is not awareness – it is structured capacity-building infrastructure that enables sustained implementation rather than one-time exposure to principles. Closing this gap would position Europe as a governance exporter, not only a regulatory model.

The AI Dialogue's unique value is its universality – every Member State has a seat. This creates three specific opportunities for advancing cooperation: First, legitimizing intermediate governance infrastructure. Current cooperation focuses on principles (UNESCO, OECD) or regulation (EU AI Act, national frameworks), creating a false binary. The Dialogue can establish education hubs, translational institutions, and capacity-building networks as a distinct governance category deserving coordinated international support. This would enable smaller states and civil society to participate not only in consultation but in sustained implementation work. Second, facilitating South-South and triangular cooperation on governance models. Current discourse centers on US-China-EU approaches, marginalizing governance innovations from Latin America, Africa, and Asia-Pacific regions. The Dialogue's inclusive structure can surface alternative models – community-led governance, indigenous data sovereignty frameworks, regional capacity-building networks – that challenge the assumption that effective governance must originate from technologically dominant states. Third, creating accountability mechanisms for soft law implementation. International frameworks proliferate but lack follow-up: who monitors whether Member States actually build capacity under OECD Principles? Who tracks whether UNESCO Recommendation commitments translate into funded programs? The Dialogue can establish peer-review mechanisms, implementation reporting, and knowledge-sharing platforms that make soft law instruments actionable rather than aspirational. Success means moving beyond declarations toward infrastructure: coordinated funding for education hubs, recognition of non-regulatory governance layers, and accountability for turning principles into practice. The Dialogue's convening power matters less than its capacity to authorize and resource the institutions that make international cooperation operational.

The AI Dialogue should connect three existing layers: Normative frameworks: The UNESCO Recommendation on the Ethics of AI (2021), OECD AI Principles (2019), and Council of Europe Framework Convention on AI and Human Rights (2024) provide essential normative consensus. The Dialogue should mandate implementation reporting – requiring Member States to demonstrate how these commitments translate into funded programs, not merely policy acknowledgment. Regional mechanisms: The EU's AI Office, African Union's Continental AI Strategy, ASEAN's governance frameworks, and Latin American regional dialogues represent diverse governance approaches. The Dialogue should facilitate peer learning among these mechanisms, enabling smaller states to adapt regulatory innovations without duplicating development costs. Mediterranean networks connecting European, North African, and Middle Eastern stakeholders offer under-utilized bridges for cross-regional exchange. Research and civil society networks: The Partnership on AI, AI Now Institute, and emerging neuroethics research networks (including European collaborations bridging neuroscience, ethics, and governance) generate evidence that should inform policy. The Dialogue should establish structured pathways for research findings to reach policymakers, and for policy questions to guide research agendas. The added value the Dialogue brings is connective tissue between these layers. Currently, normative frameworks, regional mechanisms, and research networks operate in parallel with limited coordination. The Dialogue can create structured exchange: annual reporting from Member States on UNESCO/OECD implementation, peer-review mechanisms for regional governance approaches, and researcher-practitioner-policymaker convenings that close the gap between evidence and action. This infrastructure-building role matters more than producing additional principles. The world has sufficient normative consensus – what it lacks is the institutional architecture to operationalize it.

Meaningful stakeholder contribution requires three structural changes: First, move beyond consultation-only participation. Current models invite civil society and technical communities to submit inputs, then reserve decision-making for Member States. The Dialogue should establish working groups with mixed composition – government representatives working alongside practitioners, researchers, and affected communities – tasked with producing actionable recommendations, not just position papers. Education hubs, for instance, could co-develop implementation guidelines with regulators rather than merely commenting on draft frameworks. Second, recognize contribution diversity. Not all stakeholders contribute through formal submissions. Practitioners contribute through implementation experience; affected communities through lived expertise; researchers through evidence generation. The Dialogue should create multiple contribution pathways: implementation case studies, community testimony sessions, research briefings, and peer-learning exchanges alongside traditional written inputs. Third, resource participation. Civil society organizations and smaller states lack the capacity to sustain engagement across multi-year processes without funding. The Dialogue should establish participation grants enabling consistent involvement rather than episodic consultation. Mediterranean civil society organizations, for example, could coordinate regional inputs if resourced to convene local stakeholders. Format recommendations: Annual plenary sessions supplemented by thematic working groups meeting quarterly; virtual participation options ensuring geographic inclusion; rotating locations beyond Geneva and New York to signal genuine universality; side events enabling informal peer exchange; and published implementation reports tracking how Dialogue outputs translate into Member State action. Structure matters because it determines whose knowledge counts. If only those with diplomatic infrastructure can participate consistently, the Dialogue reproduces existing power imbalances rather than correcting them.

Four constituencies remain systematically underrepresented: Practitioners implementing AI in public services – educators, healthcare workers, social service providers – who navigate governance gaps daily but rarely shape policy. Their operational knowledge reveals where frameworks succeed or fail in practice. Inclusion pathway: reserve working group seats for frontline practitioners; fund practitioner exchanges enabling peer learning; require impact assessments to incorporate implementer testimony. Communities experiencing AI harms disproportionately – workers facing algorithmic management, migrants subjected to automated border systems, marginalized groups targeted by predictive policing. Their expertise is experiential, not technical, yet essential for rights-based governance. Inclusion pathway: community testimony sessions with interpretation support; partnership with civil society organizations already serving affected populations; compensation for participation time recognizing that governance work is labor. Small and middle-income countries lacking diplomatic AI expertise. Current governance discourse is dominated by technologically advanced states and large emerging economies, marginalizing nations that will be governance-takers rather than governance-makers. Inclusion pathway: capacity-building support enabling smaller states to develop positions; South-South cooperation networks facilitating peer learning; regional hubs coordinating collective inputs. Cultural producers and creative sectors – filmmakers, musicians, writers, artists – whose work increasingly intersects with AI but whose governance concerns differ from tech sector or academic frameworks. Their contribution addresses cultural implications, linguistic justice, and the human practices that shape how AI systems are adopted. Inclusion pathway: creative industry working groups; partnership with cultural organizations; recognition that governance is not only regulatory but cultural work. Representation requires both invitation and resource – underrepresented voices need funding, interpretation, flexible formats, and decision-making authority, not merely speaking opportunities.

Five formats that move beyond traditional panel discussions: Implementation laboratories: Small groups – mixing regulators, practitioners, researchers, and affected communities – work through concrete governance challenges. Example: "How does a hospital implement the EU AI Act's transparency requirements for diagnostic systems?" Participants draft operational guidelines collaboratively, surfacing gaps between regulatory intent and institutional capacity. Outputs become implementation resources for others facing similar challenges. Peer-learning exchanges: States and organizations at similar governance maturity stages share experiences without hierarchical expertise assumptions. Example: "How are middle-income countries adapting EU regulatory models?" Participants exchange policy documents, discuss adaptation challenges, and co-develop implementation strategies. Format emphasizes horizontal learning, not expert-to-learner transmission. Scenario workshops: Participants explore governance futures through structured scenarios. Example: "What happens when algorithmic content moderation conflicts with cultural expression norms across different regions?" Groups develop divergent governance responses, then identify common principles and acceptable variation. Format surfaces value tensions that position papers obscure. Practitioner shadowing programs: Policymakers spend time with AI system implementers – teachers using educational AI, social workers managing automated eligibility systems – observing operational realities. Format builds empathy and surfaces implementation gaps invisible from regulatory offices. Cultural dialogues: Filmmakers, musicians, and artists engage policymakers on governance's cultural dimensions. Example: exhibitions showing how different cultural contexts shape human-AI interaction; performances exploring algorithmic agency; discussions on how governance frameworks enable or constrain cultural production. Format recognizes that governance is not only technical-regulatory work but shapes cultural evolution. These formats share a principle: governance knowledge is distributed across technical experts, policymakers, practitioners, and affected communities. Innovation means creating structures where that distributed knowledge becomes collective intelligence.

Our "Identity Before Tools" pedagogy demonstrates how governance education can bridge the implementation gap. The practice: Before introducing any AI system, participants identify which fundamental rights their use case affects - privacy, non-discrimination, autonomy, remedy. Only then do they evaluate technical solutions. This inverts typical AI ethics training that treats rights as post-deployment constraints rather than design prerequisites. Concrete example: A film school instructor asked how to use AI for student script feedback. Rather than immediately discussing tools, we mapped rights implications: student privacy (data retention), fairness (algorithmic bias in creative assessment), transparency (explaining AI suggestions), and intellectual property (ownership of AI-assisted work). The instructor then selected tools meeting those requirements rather than adopting the most convenient platform. Implementation mechanism: We developed workshop curricula released under Creative Commons licenses, enabling replication without dependency on proprietary platforms or consultant relationships. Organizations from education, healthcare, and public administration have adapted these materials to sector-specific contexts. This open-source approach scales governance capacity without concentrating expertise. Results: Participants report shifting from paralysis ("AI ethics is too complex for me to implement") to agency ("I can evaluate systems against rights frameworks I already understand"). Workshop demand consistently exceeds capacity, and our newsletter maintains 44-50% open rates - evidence that practitioners seek structured implementation guidance, not additional principles. The broader lesson: effective governance requires translational infrastructure connecting international frameworks to daily practice. Education hubs provide this infrastructure at lower cost than regulatory enforcement while building capacity that persists beyond compliance requirements. This approach is replicable across regions and sectors, making it suitable for international cooperation. Our Mediterranean hub model will formalize this methodology for sustained, multi-stakeholder implementation work.