POLICY BRIEF: Beyond AI Adoption: Data Sovereignty, Economic Participation and the Governance Execution Gap in Guyana and CARICOM

 

INNISS INSTITUTE FOR DIGITAL POLICY AND INTELLECTUAL PROPERTY

Beyond AI Adoption: Data Sovereignty, Economic Participation and the Governance Execution Gap in Guyana and CARICOM

POLICY BRIEF  |  NOVEMBER 2026

DOI : 10.5281/zenodo.24047419

Dr. Abiola Inniss, Ph.D., LL.M.

Scholarship for Social Change  •  Caribbean and Global South

Executive summary

Artificial intelligence is increasingly presented as a route to productivity, public-sector modernization and economic diversification. For Guyana and the Caribbean Community (CARICOM), these opportunities are genuine, but they are not self-executing. The decisive question is not whether governments and firms acquire AI systems; it is whether Caribbean institutions can govern the data, intellectual property, contractual relationships and economic value generated through their use. A region can expand AI adoption while remaining dependent on external infrastructures and suppliers, with limited influence over how its data are exploited or how benefits are distributed.

This brief applies the author’s Execution Gap diagnostic, Inniss Data Nullius framework, Digital Plantation analysis and Principle of Data Reversion to the emerging regional AI economy. It argues for a shift from technology-first announcements to sovereignty-aware, institutionally credible economic participation. It identifies five strategic policy priorities while reserving detailed diagnostic instruments and implementation designs for specialized institutional assessment. Its purpose is to inform public debate and frame actionable areas for government, regional organizations, businesses and development partners.

1. The regional opportunity is larger than technology procurement

The region is not starting from a blank slate. CARICOM approved its Vision and Roadmap for a Single ICT Space in 2017, envisaging more harmonised policy and regulatory arrangements, common frameworks, and secure technology systems. The continued relevance of that roadmap is instructive: digital integration is an institutional undertaking, not merely a connectivity project. More recently, members of CARICOM’s Council for Trade and Economic Development responsible for ICT endorsed UNESCO’s Caribbean AI Policy Roadmap in July 2026. Endorsement is a policy signal, however, and must not be confused with evidence that each jurisdiction has implemented operational safeguards, procurement disciplines or effective oversight. [1, 2] [1, 10]

AI may improve administrative efficiency, trade facilitation, agricultural forecasting, logistics, health services, education and business analytics. In Guyana, rapid economic transformation creates particular demand for capable public institutions and firms that can manage new information flows and complex commercial arrangements. Across CARICOM, shared challenges include small domestic markets, uneven technical capacity, fragmented rules, reliance on imported digital services and difficulties retaining specialist talent. These conditions differ by jurisdiction; a regional strategy cannot assume identical readiness or legal arrangements. UNESCO’s 2024 Caribbean AI Policy Roadmap identifies governance, education, culture and resilience as regional priorities. [8]

The policy debate often begins with access: connectivity, computing infrastructure, training and software. These inputs matter. Yet access does not answer who controls valuable datasets, who owns or licenses outputs, whether public institutions can audit contracted systems, or how local creators and enterprises participate in value chains. A government may digitize a service while losing bargaining power over its data and operational knowledge. A business may deploy AI while accepting contractual terms that weaken control over confidential information or intellectual property.

The economic issue is therefore one of participation, not merely consumption. Regional policy must connect investment in AI to enforceable governance, local capability formation, competitive procurement and meaningful retention of value.

Guyana illustrates the distinction between statutory foundations and operational performance. Its Data Protection Act No. 18 of 2023 provides an identifiable legislative reference point for personal-data governance. Yet the existence of legislation cannot alone establish whether agencies have sufficient trained personnel, interoperable recordkeeping, independent review mechanisms or the contractual leverage needed to supervise AI suppliers. The Act should therefore be assessed alongside actual institutional arrangements and implementation evidence, rather than treated as proof of either success or failure. [3]

2. The Execution Gap: from declared ambition to institutional performance

A useful distinction is between three layers of readiness. Normative readiness concerns the existence and clarity of applicable laws, policies and institutional mandates. Transactional readiness concerns whether purchasers and suppliers can define enforceable terms for data access, model use, intellectual property, audit rights, security and exit. Operational readiness concerns whether institutions can monitor outcomes, investigate failures, preserve records and remedy harm after deployment. Weakness in any one layer may undermine investment in the others. This layered analysis offers a public-facing account of the problem; institution-specific diagnostic tools and risk-weighting procedures require tailored professional assessment.

The Execution Gap, developed by Inniss in research presented at the 2026 European Policy for Intellectual Property (EPIP) conference, identifies a structural divide between formal policy commitments and the institutional conditions required for effective participation in AI-driven data economies. The gap cannot be closed by legislation or skills training alone. It emerges when public authorities lack coordinated mandates, procurement expertise, operational accountability, credible enforcement pathways and the ability to evaluate long-term commercial consequences. [4]

For Guyana and CARICOM, a practical diagnostic begins with four questions. First, do institutions possess legal authority and operational competence to supervise AI systems throughout their life cycle? Second, can public and private actors negotiate and enforce terms governing data access, intellectual property, security and downstream use? Third, are institutional responsibilities coordinated across trade, technology, competition, education, privacy and innovation policy? Fourth, is success measured through demonstrable economic and public outcomes rather than launches, memoranda or procurement totals?

These are not rhetorical concerns. Without clear institutional ownership, an AI initiative may be technically functional but legally opaque, financially dependent or difficult to audit. The appropriate response is not wholesale rejection of imported technologies. It is an institutional ability to choose, adapt, supervise and, where necessary, exit arrangements on terms consistent with national and regional interests.

For example, a ministry may purchase an AI-enabled records platform while agreeing to broad secondary uses of records or metadata. A small enterprise may supply locally valuable agricultural observations to a platform yet receive only a limited subscription service in return. Neither scenario automatically proves exploitation: the legality and fairness of the arrangements depend on consent, contract, applicable law, bargaining conditions and actual use. They do show why procurement price and technical performance are inadequate measures of economic participation.

3. Inniss Data Nullius and the governance of economic value

The distribution of value warrants particular scrutiny. AI value chains typically combine data collection, cleaning and annotation, computing infrastructure, model development, distribution and downstream services. Caribbean actors may participate primarily as data sources or end users while higher-margin activities, technical standards and contractual control remain elsewhere. The policy question is not whether every component must be domestically owned. It is whether regional institutions can negotiate fair terms, cultivate locally anchored services, protect legitimate rights and retain the practical ability to change suppliers. Public contracts and research partnerships are especially consequential because their terms may shape later private-market expectations.

The Inniss Data Nullius framework, advanced by Inniss as a diagnostic account of digital extraction and sovereignty deficits, identifies the risk that data generated within a society are treated in practice as readily available for external capture or monetization without adequate recognition of local governance interests and economic participation. The framework is not a claim that all data belong to the state, nor that cross-border data flows are inherently illegitimate. It asks whether legal and institutional arrangements adequately address asymmetries in access, control, bargaining power and benefit. [5]

This distinction is important in an AI economy. Training datasets, public records, cultural materials, commercial information and sector-specific data may be aggregated, transformed and reused through arrangements that are difficult for the original data-generating institutions to scrutinize. Different categories of data are governed by different legal interests, including privacy, confidentiality, copyright, database arrangements, contract, public law and community interests. Sovereignty-aware governance must respect these distinctions rather than inventing a single blanket property right in all information.

The framework also connects to Inniss’s Digital Plantation analysis: the risk of reproducing extractive economic relationships through digital infrastructures, platforms and externally controlled systems. The analogy is a structural warning, not a substitute for evidence about a particular firm or agreement. Its policy value lies in directing attention to who supplies inputs, who controls processing, who captures revenue and who retains decision-making authority. (See the author’s conceptual attribution note below.)

These frameworks are complementary rather than interchangeable. Digital Plantation names the possibility of enduring structural dependency in digital production; Inniss Data Nullius focuses attention on the treatment and capture of locally generated data; the Execution Gap examines why announced policy ambitions fail to become enforceable institutional capabilities; and the Principle of Data Reversion highlights continuing stewardship, retrieval and control over the life of an arrangement. Together they form a coherent analytical sequence: identify the structural relationship, diagnose the data-governance risk, test the institutional capacity to respond, and examine the conditions for retaining meaningful future options. None substitutes for fact-specific legal analysis.

4. The Principle of Data Reversion: retaining options and stewardship

Inniss’s Principle of Data Reversion supplies a further lens for evaluating arrangements in which information, institutional records or data-derived value move beyond the effective control of the originating institution or community. In this brief, the principle is used as a policy orientation toward continuing stewardship, recoverability and meaningful control, rather than asserted as an already-established universal legal entitlement.

Its practical relevance is evident in long-term technology contracts. Public authorities and businesses need to know whether data can be retrieved in usable formats, whether onward uses are permitted, what happens when a contract ends, how derivative datasets are handled, and whether institutional knowledge remains available after vendor transition. These questions are especially acute where a small market depends on a limited pool of external suppliers. Data reversion, properly specified through applicable law and negotiated obligations, can help frame the conditions under which institutions preserve future choices.

The governing objective is not isolation. Caribbean economies require cross-border collaboration, research access and investment. They also require credible terms of participation, including rights and responsibilities that remain intelligible after systems, vendors or public administrations change.

An economic-participation lens also changes how results are measured. Counting AI pilots, training places or software licences may show activity without revealing who earns income, who accumulates expertise, or who controls strategic assets. More revealing indicators would include local supplier participation in contracted value, retention of technical and institutional knowledge, the enforceability of licence and exit terms, the ability of regulators to obtain meaningful information, and the distribution of measurable productivity benefits. Such indicators require baselines and independent validation; the brief does not claim that these outcomes have already been measured across CARICOM.

5. Five strategic policy priorities

A. Make governance readiness a condition of major AI commitments

Before committing to high-impact deployments, public authorities should establish whether mandates, procurement oversight, information governance and independent review are adequate to the use case. The scope of assessment should be proportionate to risk, institutional capacity and the rights at stake. A published policy announcement is not a substitute for this assessment.

B. Treat public-sector data as a stewardship responsibility

Public data strategies should distinguish open information, personal information, confidential commercial material, culturally sensitive records and security-relevant datasets. Agencies need accountable custodianship, clear access conditions and safeguards for cross-border processing. The objective is responsible use and legitimate value creation, not indiscriminate restriction.

C. Link intellectual property policy to AI participation

Regional AI policy must consider copyright and licensing, trade secrets, software contracts, research commercialization, public procurement and creators’ interests together. WIPO’s guidance on generative AI and intellectual property provides a useful starting point for institutional risk awareness, but Caribbean applications require jurisdiction-specific analysis and enforceable agreements. [7]

D. Build regional coordination without erasing national differences

The regional agenda already has identifiable institutional anchors. The 2017 Single ICT Space roadmap articulates harmonisation and shared systems as objectives, while the 2026 endorsement of the Caribbean AI Policy Roadmap signals support for regional coordination. The governance question is what follows: who has responsibility for translating shared principles into domestic procedures, how cross-border expertise is pooled, and how progress is reported. A regional framework can reduce duplication, but only if national agencies retain clear authority and sufficient resources to apply it. [1, 2] [1, 2, 8, 9, 10] The CTU Caribbean AI Task Force’s 2025 interim report explicitly identifies fragmented regulation, data-sovereignty risks and a need for harmonised regional governance; these are proposed responses, not proof of implementation. [9]

CARICOM can add value through common principles, shared expertise, procurement learning and interoperable institutional approaches. Nonetheless, differences in constitutional systems, administrative capability, market structure and domestic legislation require flexible implementation. Regional coordination is most useful where it lowers transaction costs and strengthens bargaining capacity rather than creating another layer of declarations. [9, 10]

E. Measure economic participation and accountability

Indicators should move beyond numbers of trained workers or installed systems. Relevant questions include whether local firms gain viable commercial roles; whether public agencies can independently audit systems; whether contracts preserve continuity and exit options; whether creators and research institutions can negotiate appropriate terms; and whether deployment produces demonstrable improvements in public services or productivity. These are illustrative evaluation domains, not a substitute for a commissioned assessment framework.

The five priorities are intentionally strategic rather than a turnkey implementation programme. Their application will differ between a health-records system, a customs-risk model, a creative-industry licensing platform and an agricultural forecasting partnership. Each raises different combinations of privacy, intellectual property, administrative law, market access, public accountability and cross-border data issues. The Institute’s proposed advisory work would assess those combinations against the legal and institutional circumstances of a particular engagement, not presume that one imported regulatory model can be transplanted intact.

6. From policy ambition to institutional engagement

The international policy environment already offers relevant principles. UNESCO’s 2021 Recommendation on the Ethics of Artificial Intelligence addresses human oversight, accountability, data policy and institutional capacity. WIPO’s 2024 guidance identifies intellectual-property considerations for organizations adopting generative AI. Such materials are valuable reference points, but their existence does not resolve the particular institutional and economic choices facing Guyana or CARICOM. Imported principles must be translated into workable, locally accountable arrangements. [6, 7]

The next stage of the Caribbean conversation should therefore be evidence-led. Policymakers need to identify where governance responsibilities sit, which forms of dependency create material risks, what commercial interests are at stake and which institutions can implement and enforce decisions. Development partners and investors should likewise distinguish short-term adoption metrics from durable local capability and legitimate economic participation.

The Inniss Institute for Digital Policy and Intellectual Property conducts independent research and provides specialized policy advisory services on intellectual property governance, digital sovereignty, AI-driven economic participation and institutional implementation challenges. Its research frameworks support context-sensitive policy diagnostics, governance-readiness reviews, institutional analysis and strategic advisory engagements. Detailed assessment tools, implementation architecture and engagement-specific recommendations are developed within appropriately scoped professional mandates.

Conclusion

Guyana and CARICOM do not need to choose between embracing AI and protecting their economic interests. The meaningful choice is between adoption without sufficient institutional leverage and participation grounded in governance capability, legal clarity and accountable stewardship. The Execution Gap explains why ambition can fail to become capacity; Inniss Data Nullius identifies a critical risk of ungoverned extraction; the Digital Plantation lens exposes structural dependency; and the Principle of Data Reversion directs attention to continuing control and future options. Together they provide a distinct Caribbean contribution to the global discussion of AI and development. The region’s task is not simply to enter the AI economy, but to participate in shaping its terms.

References and research foundations

[1] Caribbean Community (CARICOM). (2017). Vision and Roadmap for a CARICOM Single ICT Space. https://caricom.org/documents/vision-and-roadmap-for-a-caricom-single-ict-space/

[2] United Nations in the Caribbean / UNESCO. (2026, July 7). UNESCO’s Caribbean AI Roadmap wins regional backing. https://caribbean.un.org/en/319009-unescos-caribbean-ai-roadmap-wins-regional-backing

[3] Guyana. (2023). Data Protection Act, No. 18 of 2023. Parliament of Guyana. https://parliament.gov.gy/publications/acts-of-parliament/dataprotectionact18of2023

[4] Inniss, A. A. (2026). Closing the Execution Gap: A Diagnostic Framework for Intellectual Property Participation in AI-Driven Data Economies, with Evidence from the Caribbean. EPIP 2026 conference paper; SSRN. https://doi.org/10.2139/ssrn.7451940

[5] Inniss, A. A. (2026). Inniss Data Nullius [research white paper]. Zenodo. https://doi.org/10.5281/zenodo.19699299

[6] UNESCO. (2021). Recommendation on the Ethics of Artificial Intelligence. https://www.unesco.org/en/legal-affairs/recommendation-ethics-artificial-intelligence

[7] World Intellectual Property Organization (WIPO). (2024). Generative AI: Navigating Intellectual Property. https://www.wipo.int/publications/en/details.jsp?id=4713

Conceptual attribution: Digital Plantation and the Principle of Data Reversion are original analytical concepts advanced by Abiola A. Inniss and applied in this brief. Separate bibliographic entries are not assigned without verified publication details; their definitions here are the author’s own.

[9] Caribbean Telecommunications Union (CTU), Caribbean AI Task Force. (2025). Interim Report: Toward Harmonized AI Policies and Recommendations for the Caribbean. https://ctu.int/wp-content/uploads/2025/12/CCAITF-Draft-Interim-Report-MASTER-1.pdf

[10] Caribbean Community (CARICOM). (2017). Integrated Work Plan for the CARICOM Single ICT Space (approved July 2017). https://caricom.org/documents/integrated-work-plan-for-the-cc-single-ict-space/

Conceptual attribution: Digital Plantation and the Principle of Data Reversion are analytical concepts developed by the author and applied here; this brief does not assign unverified separate publication dates or bibliographic records to them.