Pacific knowledge systems, oral, relational, held collectively across generations, are not a gap for AI to fill. They're an input current AI governance frameworks are missing entirely. Kanni Wignaraja (UN Assistant Secretary-General and UNDP Regional Director for Asia and the Pacific) and Munkhtuya Altangerel (UNDP Resident Representative, Pacific Office in Fiji) argue the region should be pushing for representation in the rooms setting AI's rules, not just support to adopt the tools once built.
By Kanni Wignaraja and Munkhtuya Altangerel
Last week, governments from across the world gathered in Geneva for the first session of the UN Global Dialogue on AI Governance, and a presentation of the inaugural report of the Independent International Scientific Panel on AI. It marked an important step towards building global rules for artificial intelligence.
While unpacking what AI can do for the Pacific, let us also consider what the Pacific can contribute to the future of AI. Both questions are relevant and connected, especially from a human development progress perspective.
The initial instinct, even anxiety, in small, remote economies could be to treat AI purely as a race from behind. And there is indeed need for that catch-up. So, the Kingdom of Tonga, for example, has moved through a national Digital Readiness Assessment, brought commercial 5G online, and is rebuilding its government services around a digital-first model, including AI tools designed to help citizens interact with the state in their own language.
Similar Digital Readiness Assessments have also been undertaken across Fiji, Samoa and Vanuatu, helping lay the foundations for AI adoption and digital transformation for the Pacific. This reflects strong national leadership and commitment across the government, with a willingness from the very top to jump into the race and speed up.
Foundational groundwork like this gives countries like Tonga real standing to speak on this issue from experience rather than aspiration. That includes investing in the AI skills and digital capabilities of people, to make the best use of the new data and services AI can offer.
However, framing the Pacific's AI challenge only as a catch-up exercise misses what the region is better placed to see than almost anywhere else. In Tonga, as across the Pacific, AI systems run into something they were not designed to encounter. Much of what local communities know about maritime navigation, weather, land and the ocean is oral, relational, and held collectively, passed down through practice and inter-generational responsibility rather than stored in a database. Some of it is sacred, and deliberately never written down.
The Pasifika Futures Report argues that the real question is not whether Pacific countries adopt AI, but whether technology supports the futures Pacific people want for themselves, especially younger generations. Digital Tuvalu offers perhaps the clearest example: a nation on the frontline of climate change using digital technology to preserve its culture, heritage and identity for generations to come, while advancing Simon Kofe's vision of a future in which Tuvaluans can connect with one another, explore their ancestry and culture, and access new opportunities for business and commerce.
This is not a barrier to AI adoption. It is an opportunity to shape how these technologies are designed and governed. Tonga is therefore well placed to bring this Pacific perspective to the future of AI, contributing views on identity, stewardship, culture and knowledge systems that remain largely absent from today's models and debates.
This brings up a larger missing piece in the current AI sphere – experiential and non-written knowledge and decisions, as well as intergenerational accountability, should not be viewed as peripheral considerations. They are critical inputs that remain largely absent from global AI governance discussions today. Ask the AI frameworks and models how they would treat a body of knowledge whose custodians choose not to digitise it, and the answer is that they were not built with this in mind. Not yet.
Tonga's own digital transformation gives it credibility to make this case precisely because it has not rushed. The Lagatoi Declaration, signed by Pacific ICT ministers in Port Moresby in 2023, set out a shared regional approach to digital transformation, infrastructure and governance long before AI dominated the conversation. Tonga's participation reflected a broader regional commitment: that a digitally connected Pacific should also be a secure and collaboratively governed one, built on shared standards rather than each nation solving the same problems alone.
The architecture exists.
What is now urgently needed is to deploy it. The Great Divergence that AI risks creating is not a distant concept. In March this year, Tonga hosted the Pacific launch of the report of that name, examining how the rapid acceleration of AI and digital technology is widening the gap between economies that build these systems and economies and societies that will simply live with the consequences.
Yet AI is already delivering results across the Pacific, from satellite and AI-enabled fisheries monitoring in Kiribati, which has helped recover around US$2 million annually in illegal fishing fines, to AI-assisted health diagnostics in countries including Nauru and Vanuatu. Digital innovation is advancing in Palau too, alongside partnerships with the European Union and UNCDF that are expanding digital services and AI-enabled financial inclusion across the region.
These are real wins, but they are wins within a system the Pacific did not design. The region already knows what it means to be on the wrong side of a divergence, in climate, in trade, in debt vulnerability. AI follows the same logic at greater speed. The difference is that the rules of AI are still being written, which means the divergence is not yet locked in. That window will not stay open for long. The time for the Pacific to shape the outcome is now, not once the frameworks are agreed and the tools are already being deployed.
The Pacific has spent decades demonstrating what it means to decide collectively, account for future generations, and treat traditional knowledge as sacred rather than extractive. These are not liabilities in the age of AI.
This is where a Pacific contribution can be pitched, while also driving for fairness in AI access and a gain in capabilities to even the odds in the AI race. Pacific representation in the forums shaping AI's future can and must emphasise the distinction this region brings to shape AI models and applications of the future.
Facts Only
* Governments from across the world gathered in Geneva for the first session of the UN Global Dialogue on AI Governance.
* An independent report on AI was presented at the event.
* The Kingdom of Tonga conducted a national Digital Readiness Assessment and brought commercial 5G online.
* Digital Readiness Assessments were undertaken in Fiji, Samoa, and Vanuatu.
* Local communities hold knowledge about maritime navigation, weather, land, and the ocean through oral, relational practices.
* Some local knowledge is considered sacred and is not written down.
* The Pasifika Futures Report questions whether technology supports desired futures for Pacific people.
* Digital Tuvalu uses digital technology to preserve culture and heritage.
* Experiential and non-written knowledge and intergenerational accountability are noted as critical inputs absent from current AI governance discussions.
* Pacific ICT ministers signed the Lagatoi Declaration in Port Moresby in 2023 regarding digital transformation.
* Tonga hosted the Pacific launch of a report on the Great Divergence between AI-building economies and living with consequences.
Executive Summary
Governments from around the world gathered in Geneva for the first session of the UN Global Dialogue on AI Governance and a presentation of an independent report on AI. The discussion centered on whether Pacific knowledge systems, which are oral, relational, and held collectively across generations, represent a gap that AI needs to fill, or if they offer unique contributions. Speakers argued that the region should focus on representation in setting AI rules rather than just adopting tools.
The experience of small, remote economies has involved national digital readiness assessments across nations like Tonga, Fiji, Samoa, and Vanuatu, which have supported the foundation for AI adoption. While addressing AI as a catch-up exercise is one view, another perspective emphasizes that local knowledge regarding maritime navigation, weather, and ocean systems exists in non-digitized, collective forms that challenge current AI models.
The argument suggests that the focus should be on ensuring that AI development supports the specific futures desired by Pacific peoples, such as preserving cultural identity and exploring ancestry, rather than viewing this as a mere technical adoption issue. This involves recognizing experiential and non-written knowledge and intergenerational accountability as critical inputs for governance.
Full Take
The core tension in this discourse lies between the imperative for rapid technological catch-up and the necessity of respecting epistemic frameworks that predate digital data structures. The narrative pivots away from framing the Pacific experience solely as a deficit requiring remediation and toward positioning it as a source of novel governance input. The unspoken assumption being challenged is that knowledge, if not digitized, holds no value in the context of global technological advancement; instead, this text posits that the manner in which AI systems are built—and thus what they govern—is fundamentally shaped by the data and accountability structures prioritized by their custodians.
The assertion that experiential and non-written knowledge should not be peripheral is a direct challenge to data-centric models prevalent in contemporary machine learning. The implication for AI governance is profound: if frameworks ignore embodied, relational knowledge, they risk creating systems that are brittle or exclusionary when applied to diverse human realities. The invocation of the "Great Divergence" links regional socio-economic vulnerability (climate, debt) directly to technological divergence, suggesting that ignoring localized context exacerbates systemic inequalities. This suggests a pattern where global technical frameworks often operate on an abstracted reality, which inevitably marginalizes those whose knowledge systems are not easily quantified or indexed—a form of epistemic colonialism where the cost of adopting new tools is assumed to be neutral. The urgency lies in seizing the opportunity presented by digital transformation not merely for adoption, but for asserting a unique ethical and cultural modality within the architecture of future intelligence systems.
Bridge questions: If knowledge systems are inherently relational and non-extractive, what concrete mechanisms can be developed to integrate intergenerational accountability directly into algorithmic oversight? How can global AI governance structures be redesigned to value and operationalize custodianship over data, rather than merely collecting it? What structural changes must occur to shift the locus of power in AI development from the builders to those who hold experiential knowledge?
Sentinel — Human
The text presents a coherent argument positioning Pacific knowledge systems as essential inputs for AI governance, using national digital transformations as evidence to advocate for region-specific representation in global AI rule-making.