University of Arts, Media and Communication

To begin with, I expect the dialogue to produce a governance framework. This would enable institutions to implement AI governance tailored to their needs, starting with basic ethical principles such as transparency, accountability, and human oversight, and then advancing to more developed regulatory systems. This ensures that no institution is disadvantaged due to resource limitations. Second, there must standard for ethical use. These should go beyond broad principles to include concrete requirements such as disclosure of AI use, rules on authorship and attribution, and safeguards for data privacy. Ethical use should be the foundation of governance, not an afterthought. Third, the dialogue must facilitate the establishment of a global-local translation paradigm. Most existing models have been created in environments of high resource availability. A successful outcome must ensure room for the adaptation and development of policy templates and platforms that can help countries such as Ghana align global standards with their local needs. Fourth, there must be a commitment to capacity building. This includes training for educators and regulators, as well as developing AI literacy among students. Fifth, the dialogue should recognize the central role of data governance. Support for building structured datasets and digitizing institutional processes is essential for effective AI governance in developing contexts. Lastly, success would mean that countries like Ghana leave with usable outputs: draft institutional policies, implementation roadmaps, and partnerships for technical support. In essence, the dialogue would largely succeed if it moves from principles to practice and produces frameworks that can be implemented within real institutional constraints.

• AI capacity-building
• Open-source software, open data and open AI models
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Interoperability of governance approaches

These four areas I have selected represent a prioritization that is closely aligned with the needs of developing institutional contexts such as Ghana, where AI development is outpacing governance development. AI Capacity Building is the most immediate priority. Effective governance depends on the ability to govern institutions that can effectively use and regulate AI. There is a need to train universities, regulators, and policymakers on effective and ethical use of AI. Without such a background, effective governance is impossible. The social, economic, ethical, cultural, linguistic, and technical implications of AI are equally critical. It is imperative to realize that AI is not value-free. It can be biased, discriminatory, and culturally insensitive. In a context such as Ghana, there is a need to ensure that AI is inclusive, culturally sensitive, and ethical. In a university context, questions of integrity, authorship, and fairness are critical. The interoperability of governance frameworks is also significant because AI works across national boundaries, and governance frameworks are fragmented. There is also a need for alignment between global and national frameworks. For instance, in the context of Ghana, it is necessary to adapt global frameworks in ways that are compatible with our local laws, institutions, and educational frameworks. Lastly, open-source software, open data, and open AI models are also significant because many institutions in developing contexts are resource-constrained and lack access to proprietary software and other relevant platforms. Open-source platforms offer opportunities for innovation and local adaptation. They are also necessary for generating context-specific data, which is required to build context-specific AI models. These are significant because they offer an ethical, inclusive, and implementable framework for AI governance.

Yes. One important cross-cutting issue is data scarcity and data quality. In many developing contexts, such as Ghana, limited digitization and fragmented records constrain both AI development and effective governance. Without reliable local datasets, AI systems may be inaccurate or biased. Another emerging issue is AI literacy and ethical l use. Governance is not only about rules but about how students, researchers, and professionals actually use AI tools in practice. Finally, there is a growing need to protect knowledge systems from overreliance on AI-generated content that may dilute originality, critical thinking, and scholarly rigour.

Governance and ethical usage gaps are already visible in Ghana's higher education and policy space. The rapid uptake of generative AI is outpacing institutional rules, leaving universities without clear guidance on ethical use, authorship, and assessment. This creates risks of academic misconduct, uneven practices, and uncertainty among students and faculty. Limited AI capabilities and digital infrastructure hinder effective oversight, while data limitations and low levels of digitization hinder the development of context-specific AI systems and effective governance. However, these gaps in knowledge and infrastructure create opportunities for development. For instance, Ghana can adopt flexible, context-specific models, invest in AI literacy, and leverage open-source technologies to minimize costs. There is an opportunity to integrate Ghana's AI strategy with higher education reforms, positioning higher education institutions as hubs for ethical AI use and innovation in AI governance.

The AI Dialogue may act as a bridge between global principles and their implementation. I believe it may foster a common set of principles for the ethical use of AI, while still allowing scope for adaptation across countries. It may build stronger partnerships in capacity building, pairing countries with technical expertise in this area with those in the process of building their systems of governance. It may improve interoperability in the systems of governance, so that they are aligned at the global level. It may also establish a platform that ensures the participation of developing countries in building AI governance, rather than adopting models already in place.

The AI Dialogue should leverage existing initiatives, including UNESCO's AI ethics framework, the African Union's AI strategy, and national AI strategies such as Ghana's National AI Strategy. It should also leverage university-based initiatives, research networks, and emerging public-private partnerships in AI-related capacity building and data governance. The AI Dialogue's value proposition is in coordination and translation. It could bring order to the current chaotic, duplicative efforts in AI governance and bridge the gap between global standards and local realities. It could also ensure that developing nations not only follow AI governance standards but also help shape them.

There are various stakeholders that need to contribute in their own way. The government may contribute through policy and regulatory expertise. The academic sector and researchers may contribute through evidence and pilot work. The private sector may contribute technical expertise and tools, and civil society may contribute ethics, inclusion, and accountability. The Dialogue may be designed in the following manner: • Thematic Working Groups: on key areas like ethics, capacity, and data governance • Regional Consultations: to ensure a sense of place and ownership • Implementation Tracks: that deliver tangible outputs like policy templates and roadmaps A light reporting mechanism may be implemented to monitor and sustain engagement.

The voices of the Global South, particularly policymakers, educators, and researchers, are also underrepresented. Informal sector participants, small organizations, and non-English speaking groups are also excluded. Moreover, students and junior researchers, as the main consumers of AI tools in education, have limited opportunities to contribute. Inclusion, as mentioned, demands careful design. It includes the funding of participants from low-resource countries, multilingual engagement, and the development of regional hubs that connect to the global conversation. It also includes the ability of universities and local research centers to participate. In addition, student/practitioner discussions must be integrated.

I suggest using formats that move beyond speeches to active problem-solving. We can deploy • Policy Labs where mixed groups co-design AI governance templates for real contexts. • Case-based simulations on issues like academic integrity or data misuse • Regional breakout sessions to surface local realities and solutions • Live peer review clinics where institutions critique draft policies • Showcase sessions for practical tools, open-source models, and pilot projects These formats, I submit, will encourage collaboration, practical outputs, and learning across contexts.

There are also emerging practices that can provide us with practical examples. For instance, some prominent institutions, such as Oxford and Harvard, have adopted disclosure and attribution practices that require users to disclose their use of AI and remain accountable. Similarly, Monash University incorporates AI into research integrity frameworks, and the University of Sydney incorporates assessment redesign, allowing the controlled use of AI with disclosure. The UNESCO framework for AI ethics promotes human oversight, fairness, and accountability, and the OECD AI principles promote interoperable governance. Open-source platforms and models can enable accessible innovation, especially in resource-constrained settings. In addition, AI literacy programs for students and faculty have been found to be critical, as they need to understand the benefits and risks of AI in practical contexts.