Autorité de régulation des postes et télécommunications

It must move beyond end-user software regulation to address the global economic and physical realities of AI. A successful dialogue must yield the following actionable outcomes: 1. ​A Value-Chain Centric Framework: The dialogue must define rules for a new productive system based on data, compute, and algorithms. Success means establishing mechanisms that allow developing countries to position themselves equitably across all segments, from physical infrastructure to applications. Failing to do so risks consolidating structural technological and economic dependencies. 2. Material Ethics and Traceability Standards: Responsible AI begins directly at its material foundations. A successful outcome requires adopting international standards for the traceability of critical minerals and energy sustainability. The governance of AI must connect the physical extraction of resources directly to algorithmic ethics. 3. Equitable Infrastructure Access: AI computing capacity is a major determinant of economic competitiveness. The dialogue must establish international mechanisms fostering access to shared infrastructures, open models, and local capacity building. Open standards must be recognized as strategic instruments to drastically reduce entry barriers. 4. Adaptive Regulatory Mechanisms: Regulation must act as a lever to accelerate innovation. We need the endorsement of contextualized frameworks, such as hybrid regulatory sandboxes, allowing emerging economies to test solutions in real conditions while strictly guaranteeing digital sovereignty and data security.

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

As a regulator in the Democratic Republic of the Congo, these selections reflect our urgent need to utilize AI as a lever for global balance and equitable economic transformation. If these areas are ignored, current imbalances will amplify, cementing structural technological and economic dependencies for the Global South. 1. AI Capacity-Building: The DRC possesses a rapidly growing youth demographic and vast human capital potential. To transform this demographic dynamic into an economic opportunity, we must invest heavily in local skills, STEM education, and vocational training. Building local capacity is essential so that AI acts as an amplifier of human capabilities and a driver of job creation rather than a displacer of labor. 2. Social, economic, ethical, cultural, linguistic and technical implications of AI: For the DRC, AI's implications span its entire physical and digital lifecycle. Ethically, responsible AI begins at its material foundations, specifically, the extraction of our critical minerals and energy resources. Culturally and socially, we must prioritize solutions adapted to local realities, ensuring the integration of national languages and local content to prevent a widening digital divide. Economically, AI must serve to formalize our economy and improve productivity. 3. Interoperability of governance approaches: Emerging economies require adaptive, contextualized regulatory frameworks, such as hybrid regulatory sandboxes, that manage risks without stifling technological dynamics. Interoperable governance ensures that our domestic rules integrate seamlessly with global open standards, allowing the DRC equitable participation in the global AI value chain.

Yes, two critical cross-cutting issues remain largely uncaptured by the standard themes: the material foundations of AI and structural economic dependency within the AI value chain. First, current frameworks treat AI as purely digital, ignoring its massive physical footprint. As the Democratic Republic of the Congo, we sit directly at the intersection of the material foundations and economic uses of artificial intelligence. An enormous blind spot in global governance is the disconnect between algorithmic ethics and the physical extraction of the critical minerals and energy required to power AI data centers. "Responsible AI" must begin at its material foundations. We urgently need international standards that enforce the traceability of critical minerals and energy sustainability as core components of AI governance, not just environmental policy. Second, the current discourse focuses heavily on regulating end-user applications rather than addressing the conditions of creation, access, and global value capture. Without actively anchoring AI governance in the global value chain, we risk reproducing and amplifying current global imbalances. Developing nations must be empowered to position themselves equitably across multiple segments of this chain, from physical infrastructure and data to applications. Otherwise, the AI revolution will merely consolidate structural technological and economic dependencies for the Global South.

As a regulator in the Democratic Republic of the Congo, I observe that current global AI governance gaps pose severe systemic risks, while recent advances offer unprecedented transformational opportunities for our region. The most critical gap is the failure to regulate AI as a complete global value chain. For the DRC, this lack of structural governance risks reproducing and amplifying existing imbalances, thereby consolidating long-term technological and economic dependencies. While we possess the critical natural resources that physically build AI, current frameworks fail to link algorithmic ethics to the traceability and sustainability of these material foundations. Furthermore, without international mechanisms guaranteeing equitable access to computational infrastructure and open models, the economic divide between nations will irreparably widen. Conversely, targeted advances in AI present a structural lever to transform the DRC's rapid demographic growth into a decisive competitive advantage. By aggressively pursuing AI capacity-building, we can utilize AI as an assistive technology that amplifies human capabilities, accelerates the formalization of our economy, and creates sustainable employment. Additionally, developments in flexible governance. such as the deployment of hybrid regulatory sandboxes, allow emerging economies to test localized solutions in real-world conditions, safeguarding our digital sovereignty without suffocating technological innovation. Ultimately, bridging these gaps is essential for the DRC to secure an equitable position across the entire AI value chain, moving beyond mere resource extraction.

The DRC is on of the physical engine of the global AI revolution. We provide the critical minerals essential for computing hardware and hold immense clean energy potential, alongside a massive, dynamic youth population. Yet, we remain trapped at the bottom of the AI value chain, supplying the raw materials while the technological and economic benefits are captured almost entirely by the Global North. The true added value of the AI Dialogue must be to serve as the historic turning point that corrects this paradox. The Dialogue must not be just another forum for software regulation; it must be the mechanism that translates the Global South's physical and demographic wealth into digital equity. To achieve this, the Dialogue must build upon and actively connect with: ​The UN Panel on Critical Energy Transition Minerals: To formally link the physical extraction of AI's hardware to the digital governance of its models. "Responsible AI" must mandate fair economic returns and technology transfer for the countries supplying its physical foundations. The Dialogue's ultimate value is to stop treating countries like the DRC as mere resource quarries, and instead integrate us as equal partners in AI innovation and value creation.

To move beyond abstract declarations, the AI Dialogue must directly integrate with the following specific initiatives: ​1. The UN Secretary-General's Panel on Critical Energy Transition Minerals: The DRC insists that AI ethics must cover its entire lifecycle, starting with its material foundations. The Dialogue must operationalize this Panel's guidelines to enforce international standards for the traceability of critical minerals and energy sustainability. 2. The African Union Continental AI Strategy: To establish an operational balance between regulation and innovation, the Dialogue must adopt the AU's contextualized approach. Specifically, it should connect with regional bodies to fund and standardize "hybrid regulatory sandboxes". This allows emerging economies to test solutions in real conditions while guaranteeing digital sovereignty and data security. ​3. The Global Digital Compact & Digital Public Goods Alliance: The Dialogue must build on these mechanisms to mandate equitable access to shared computational infrastructures and open-source models. Open standards must be enforced as strategic instruments to reduce entry barriers for local developers. The unique added value of this Dialogue is establishing an explicit, binding link across the entire AI value chain. Its purpose is to implement mechanisms that actively prevent the reproduction of structural technological dependencies, ensuring industrializing nations capture equitable value from infrastructure to application.

To avoid becoming another rhetorical echo chamber, the AI Dialogue must discard the traditional format of broad diplomatic plenaries. It must be structured as an operational negotiation forum that treats AI as an industrial ecosystem. ​Recommended Structure: ​Value-Chain Working Groups: The Dialogue must abandon generic thematic panels and instead divide into specific tracks reflecting the AI supply chain: 1) Material & Energy Infrastructure, 2) Data & Compute Access, and 3) Applications & Governance. Structuring the dialogue by value chain ensures industrializing countries can position themselves equitably across specific segments. ​Decentralized, Reality-Based Hosting: Rotate technical sessions outside of New York and Geneva. Hosting the Material & Energy Infrastructure track in a resource hub like in Kinshasa would force global tech stakeholders to directly confront the physical, ecological, and energetic realities underlying the algorithms they regulate. How Stakeholders Must Contribute: ​Global Tech Corporations: Must move beyond lobbying for software "safety" and actively contribute to the Data & Compute track by committing IP to open standards and providing access to shared computing infrastructures. Resource-Rich Governments: Should co-chair the infrastructure tracks, leveraging their critical minerals and demographic dynamics to negotiate equitable integration into the global digital economy, ensuring they are not just raw material providers. Local Innovators & Civil Society: Must be embedded in technical tracks to mandate the inclusion of national languages and local cultural realities, preventing the homogenization of global AI models.

The most critically underrepresented voices in global AI governance are the communities that form the physical and demographic foundation of the AI ecosystem: resource-supplying nations, non-Western linguistic groups, and the emerging workforces of the Global South. Underrepresented Perspectives: ​The Material Base: Nations supplying the critical minerals and energy required to physically build AI. Their populations are entirely excluded from digital governance, creating a fatal disconnect between algorithmic ethics and the ecological/economic realities of extraction. ​Linguistically and Culturally Diverse Populations: Current global models are heavily centralized around Western data and languages. Communities reliant on local and national languages are structurally excluded, resulting in a homogenized AI that lacks local relevance and accelerates a cultural digital divide. ​The Emerging Workforce: In industrializing economies with high demographic growth, global governance often views AI merely through the lens of risk and job displacement. This ignores the pressing perspective of local workforces who urgently need AI as an accessible tool for economic formalization, capacity building, and productivity. Mechanisms for Inclusion: ​Value-Chain Representation: Resource-rich nations must be elevated from passive observers to active co-designers. They must hold institutional seats to negotiate technology transfer and equitable infrastructural access directly. ​Mandating Linguistic Equity: The Dialogue must structurally support and fund the integration of national languages and local data into open-source foundation models, treating local cultural representation as a strategic priority, not an afterthought. ​Empowering Local Ecosystems: Governance must actively integrate grassroots tech hubs, STEM educators, and local entrepreneurs to ensure regulations foster, rather than suffocate, localized technological innovation.

To effectively foster dynamic engagement and break the inertia of traditional diplomatic plenaries, the AI Dialogue must adopt operational formats that reflect AI's technical reality and global physical footprint. ​1. Live "Hybrid Sandbox" Simulations: Move beyond theoretical policy debates by conducting live, cross-border regulatory simulations. Policymakers, developers, and local innovators should collaboratively stress-test proposed global regulations against real-world constraints typical of emerging economies. Testing rules against localized data scarcity or connectivity limits enforces pragmatism and validates true governance interoperability. ​2. Value-Chain Reverse Engineering Sessions: Implement interactive mapping workshops that trace foundation models backward from end-user applications to the data centers and the critical minerals powering them. Forcing stakeholders to visually and economically deconstruct the technology ensures the Dialogue confronts the material foundations of AI, making structural dependencies undeniable. ​3. Governance Infrastructure "Hackathons": Instead of passive panels, host technical working groups where Global North tech leaders and Global South STEM talent collaborate to solve specific governance bottlenecks. Tasking groups to prototype open-source oversight tools or frameworks for local linguistic integration forces practical capacity-building and demonstrates how regulation can accelerate innovation. ​4. "Reversed Pitch" Reality Checks: Format sessions where tech corporations listen to regulators from industrializing nations pitch specific national challenges (e.g., public sector deployment, energy sustainability). Companies must then propose on-the-spot, open-standard solutions. ​These formats transform the Dialogue from a rhetorical forum into an actionable, value-chain-centric negotiation.

To translate principles into operational realities, effective AI governance must focus on adaptive frameworks, physical traceability, and shared infrastructure. The following concrete approaches offer immediate, structural solutions: 1. Hybrid Regulatory Sandboxes (The Innovation-Regulation Balance): Static, rigid regulation suffocates emerging markets. A concrete solution is the deployment of "Hybrid Regulatory Sandboxes," a strategic priority for the DRC. These mechanisms combine technical supervision with legal flexibility, allowing developers to test AI solutions in real-world conditions while simultaneously ensuring digital sovereignty, data security, and risk mitigation. 2. Material Traceability Protocols (Hardware as AI Ethics): AI governance currently suffers from a purely digital bias. A necessary practice is integrating supply chain audits into algorithmic compliance. "Responsible AI" must mandate the verifiable traceability of critical minerals and energy sustainability. This practice formally connects the ethical extraction of physical resources in the Global South directly to the governance of algorithms. 3. Sovereign Compute & Shared Digital Public Infrastructure (DPI): To break structural economic dependencies, policies must support the creation of shared international computing infrastructures and enforce open-source interoperability. By pooling computational resources, we allow developing nations to actively participate in the AI value chain. Open standards must be enforced as strategic tools to drastically lower entry barriers. 4. Algorithmic Localization Mandates: A concrete practice for digital inclusion is mandating the integration of national languages and local cultural datasets before foundation models are deployed in public sectors.