Ministry of Foreign Affairs - Brazil

A successful first Global Dialogue on AI Governance would be defined by its ability to shift the international focus from purely commercial or security-based interests toward a development-oriented, inclusive, and human-centric framework. It would be welcome, as well, to have effective, practical knowledge produced and distributed among its participants, to ensure that ideas and conclusions drawn from the experience are reflected in policy actions.

• Safe, secure and trustworthy AI
• AI capacity-building
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Transparency, accountability, and human oversight

Safe, Secure and Trustworthy AI: For Brazil, safe, secure and trustworthy AI is a priority because it allows societies to maximize the social, economic and environmental benefits of AI while mitigating risks related to discrimination, surveillance, disinformation, concentration of economic power and misuse of data. At the same time, this agenda must not be reduced to the securitization of AI. It should preserve the centrality of development, digital inclusion, regulatory sovereignty, information integrity, human rights and the meaningful participation of the Global South in AI governance. AI capacity-building: Addressing the digital divide is a strategic imperative to ensure that the Global South moves from being primarily a provider of raw data to becoming a producer of science, technology and innovation. Brazil advocates for international cooperation that includes technology transfer, infrastructure support, skills development and access to quality data and computing capacity. Such cooperation should enable countries to build and deploy AI systems that respond to their own development needs, priorities and institutional capabilities. Social, economic, ethical, cultural, linguistic and technical implications of AI: AI must be oriented toward reducing poverty and inequality and advancing sustainable development, while fully respecting human rights. Governance frameworks should ensure that AI systems respect linguistic, cultural, racial, geographic and demographic diversity; mitigate algorithmic bias; promote information integrity, tolerance and respect in the digital space; protect the integrity of democratic processes; and safeguard labor dignity in the face of the structural impacts of automation. AI should expand human capabilities rather than deepen existing inequalities. Transparency, accountability and human oversight: For Brazil, transparency, accountability and human oversight are essential conditions for AI systems to be trustworthy, auditable and compatible with human rights. AI systems should allow a meaningful understanding of their operation and impacts, provide clear mechanisms for accountability, and preserve final human control, particularly in sensitive applications. This priority helps reduce risks, prevent discrimination and ensure that automated decisions do not replace political, legal and ethical responsibility.

1. Environmental Sustainability and "Green AI" One of the most prominent emerging issues is the intersection of AI and climate change. While AI can be a tool for climate mitigation, its development generates a significant environmental liability due to the high energy and water consumption of data centers. Brazil emphasizes that AI governance must ensure the technology minimizes its own environmental footprint. Furthermore, Brazil views its renewable energy matrix as a comparative advantage to lead in the creation of green digital infrastructure, a topic that links innovation directly to environmental policy. . 2. Economic Justice and Fair Remuneration While "open data" is listed, Brazil identifies a deeper cross-cutting issue: the extractive nature of the current data economy. There is an asymmetry in the AI ecosystem where individuals and governments provide data for currently have their data freely used by transnational corporations, while these corporations capture all the resulting value. This provides an opportunity for the Dialogue to seek advances in governance that enable countries to fully enjoy the economic benefits from the data economy. 3. Impact on the World of Work and Labor Dignity Brazil places an emphasis on the structural impact of AI on the labor market. Beyond general social implications, there is an urgent need to protect labor conditions, collective bargaining, and social security in the face of rapid automation. The goal is to ensure that AI does not lead to the extinction of human functions but serves to strengthen the dignity and well-being of workers. 4. Information integrity AI systems now mediate how information is created, ranked, recommended, translated, and monetised across the platforms on which billions of people depend. This affects every other thematic area - from the safety of AI systems, to human rights, to capacity-building, to the equitable distribution of opportunities. The integrity of the global information ecosystem is under increasing strain, and AI deployment is intensifying that strain. The UN Global Principles for Information Integrity (2024) provide the agreed multilateral framework for addressing these risks, and the Belém Declaration on Information Integrity on Climate Change demonstrates how it can be applied to a specific global priority. The UN Global Risk Report 2024 identified mis- and dis-information as the top current global vulnerability, a major risk for which multilateral institutions are not sufficiently prepared. Likewise, the World Economic Forum's Global Risks Report 2026 identified mis- and disinformation as the second most severe risk in the short term and the fourth in the long-term - right after three environment-related risks.

1) Safe, Secure and Trustworthy AI. Governance gaps affect Brazil by exposing society to AI systems that may deepen discrimination, enable surveillance, spread disinformation, intensify concentration of economic power and misuse personal or strategic data. The main challenge is to avoid the securitization of AI: safety cannot be reduced to national security concerns defined by a few advanced economies. For Brazil, "safe and trustworthy" must also mean development-oriented, inclusive, rights-respecting and compatible with regulatory sovereignty. The opportunity is to shape a broader global agenda that links safety to information integrity, digital inclusion and the participation of the Global South. 2) Capacity-building. Governance gaps reinforce asymmetries between countries that control data, computing power, infrastructure and frontier models, and countries that mainly provide raw data or consumer markets. For Brazil, the challenge is to avoid technological dependency and exclusion from AI value chains. The opportunity is to promote cooperation for technology transfer, infrastructure, skills, open innovation, quality data and computing capacity, enabling Brazil and the Global South to become producers of AI systems aligned with their own priorities. 3) Social, Economic, Ethical, Cultural, Linguistic, and Technical Implications. Challenges: AI systems can reinforce negative biases that harm vulnerable groups and hinder human rights. The rise of generative AI and deepfakes pose a threat to information integrity and democratic institutions. Digital platform's AI-driven recommender systems and advertising tools often spread and monetize disinformation. Opportunities: Under the "Data for Development" paradigm, Brazil views AI as a unique strategic opportunity to accelerate solutions in health, education, and agriculture, specifically aimed at reducing poverty and inequality. 4) Transparency, accountability and human oversight. Brazil is affected by the absence of clear international standards on auditability, explainability, liability and final human control over automated decisions. This creates risks in sensitive areas such as public services, health, education, labor, access to rights and democratic processes. The main challenge is to ensure that automated systems do not replace political, legal and ethical responsibility. The opportunity is to advance governance models based on transparency, independent auditing, due diligence, accountability across the AI value chain and meaningful human oversight.

The Global Dialogue on AI Governance can be a central mechanism to foster inclusive, legitimate, and effective international cooperation through the following functions: 1. Centralizing Governance within the United Nations The primary role of the Dialogue can be to streamline the current fragmented initiatives on AI governance. The United Nations, as the most legitimate, universally inclusive, and representative forum for global decisions on AI, can provide a formal process to ensure that governance is not dictated by exclusive "clubs" of powerful nations. 2. Empowering the Global South The Dialogue may serve as a platform to ensure that countries of the Global South are not mere spectators of technological advancement but active participants in decision-making processes. 3. Promoting the "Data for Development" Paradigm The Dialogue is a space to advance a new global consensus that treats data as a strategic collective asset for the public good. International cooperation through this forum could focus on: Aligning AI development with the Sustainable Development Goals (SDGs) and poverty reduction Moving beyond purely commercial views of data toward a framework that supports social and economic justice. Establishing mechanisms for fair and transparent remuneration for the citizens and governments that provide the data used to train global AI models. 4. Fostering Interoperability and Open Innovation The Dialogue may encourage cooperation on technical standards and interoperability that are non-discriminatory. It may promote open-source "foundation" models and "Science Openness" to prevent "policy capture" by private actors through proprietary systems, ensuring that the benefits of AI are accessible to all nations regardless of their development level.

There is a wide array of existing efforts that the AI Dialogue should connect with to avoid redundancy and erosion of the multilateral system: UN System and Specialized Agencies: The Dialogue should draw from the UNESCO Recommendation on the Ethics of Artificial Intelligence, , the ITU's technical expertise, UNCTAD's work on the digital economy, and the UNFCCC's #AI4ClimateAction initiative. It also connects directly to the Global Digital Compact and the UN Global Principles for Information Integrity. The World Intellectual Property Organization may also offer important insights on AI, especially on mechanisms for fair and transparent remuneration. Intergovernmental Blocs: It should link with the acquis of the G20 Digital Economy Working Group (DEWG) and its Task Force on AI and Innovation, as well as the BRICS Leaders' Declaration on AI Governance. Specialized and Regional Forums: This includes the Global Partnership on AI (GPAI), now integrating with the OECD, and regional bodies like the Latin American AI Working Group, eLAC (Digital Strategy for Latin America and the Caribbean), and the Mercosur action plan on digital government. Technical Standards Bodies: Collaboration with the ISO, IEC, and IEEE is essential to align technical parameters and protocols with broader policy goals. Bilateral Frameworks: Existing cooperation agreements, such as those Brazil has signed with China, Chile, and Ecuador, provide templates for sharing computational infrastructure and developing "foundation" models. The AI Dialogue brings specific value in its universal legitimacy and inclusion that other, more exclusive forums lack; therefore, it would bring more added value as a forum to seek harmonization and convergence among governance and technical standards. It could encourage the correction of global asymmetries while also reinforcing the right of States to regulate their own digital economies, ensuring that international cooperation respects national digital sovereignty.

Brazil advocates for an approach where various actors provide unique contributions based on their respective roles and competencies. While governments hold a fundamental and priority role in leading discussions on AI governance and implementing internal regulations within their jurisdictions, the private Sector, civil society, and international organizations, should work in partnership with governments, providing exclusive knowledge, perspectives, and resources to help build a comprehensive governance model. Technical and academic communities should also contribute with specialized expertise to ensure that governance is informed by scientific and engineering foundations. Brazil also expects the International Scientific Panel on AI, as an independent body, to provide evidence-based insights to support the dialogue's objectives. Following the Tunis Agenda and the Global Digital Compact, actors should participate according to their specific functions and responsibilities, keeping the process human-centered and development-oriented, while maintaining governments as the leading actors in decision-making. The format should ensure that the decision-making process respects linguistic, cultural, and demographic diversity, which is essential for a truly global and equitable dialogue.

The discussion on global AI governance is currently marked by significant asymmetries, leaving several key groups and perspectives marginalized. The Global South and Developing Nations: There is an increasing "digital divide" between technologically advanced nations and those still developing basic digital infrastructure and computing power. Currently, the regulatory landscape is dominated by the approaches of major economic powers, with limited representation of the perspectives and needs of developing countries. Research is also concentrated in the Global North. By consolidating UN-Centered Multilateralism, the United Nations and its specialized agencies (such as UNESCO, UNCTAD, and ITU) must be utilized as the most universally inclusive and representative forums for decision-making, preventing governance from being restricted to exclusive "clubs" of powerful nations. Vulnerable and Marginalized Social Groups: AI systems often lack the input of women, minorities, and groups in vulnerable situations, such as children, afro-descendants, LGBTQIA+ individuals, indigenous peoples, and the elderly. Without their active participation, AI risks reinforcing negative biases and exacerbating racism and sexism. Data Providers (Citizens, Universities, and Local Governments): While individuals and institutions provide the vast scale of data required to train AI, they often do so for free, while large transnational platforms capture all the economic value. This data also commonly flows from the Global South to the Global North, the latter capturing its value, in a flow that deepens the "digital divide". International rules should include them by establishing mechanisms for "fair and transparent remuneration" for data providers, therefore correcting global asymmetries and ensuring that the wealth generated by AI is equitably distributed back to the communities that provided the input.

Clear and short political statements in plenary The plenary should provide strategic orientation for the Dialogue. Statements by Member States and other high-level participants should be concise, focused, and political in nature. They should identify shared priorities, such as inclusive AI governance, capacity-building, human rights, safety, interoperability, and development-oriented cooperation. The objective should not be to deliver long national statements, but to establish a common direction for the work of the Dialogue. Thematic, solution-oriented multistakeholder breakout sessions The core of the Dialogue should be organized around thematic breakout sessions with participation mainly from governments, in order to reinforce the intergovernmental character of multilateral negotiations around AI governance. These sessions should be practical and solution-oriented. They could address issues such as AI capacity-building, safe and trustworthy AI, governance interoperability, human rights, transparency and accountability, open-source AI, information integrity and AI, access to data and compute, and the social, economic, cultural and linguistic implications of AI. Each session should be designed to identify concrete challenges, good practices, possible cooperation initiatives, and areas where further international coordination is needed. A follow-up process with clear next steps The Dialogue should not be a one-off exchange of views. It should establish a follow-up process with clear steps stemming from the discussions and decisions undertaken during the Dialogue. This could include a Chair's summary, identification of priority workstreams, requests for further inputs, mapping of existing initiatives, voluntary partnerships, and a timeline for reporting back at the next session. This would help ensure continuity, accountability, and practical value. In this format, stakeholders would contribute according to their respective roles: Member States would provide political guidance; international organizations would connect the Dialogue to existing mandates; technical bodies and academia would provide evidence and expertise; the private sector would share implementation experience and resources; and civil society and affected communities would help ensure that governance remains inclusive, rights-based and responsive to real-world impacts.

Effective AI governance is promoted through a combination of national strategic planning, robust digital infrastructures, and international ethical frameworks. 1. National Policies and Legislative Frameworks Plano Brasileiro de Inteligência Artificial (PBIA): A comprehensive strategy that systematizes international components across all its guidelines to foster local capacity and strategic development. Brazil's AI Bill (PL 2338/2423): A specific legislative project currently under discussion in the Brazilian Congress aimed at the formal regulation of AI in the country. Bill 6237/2025, presented by the President in December 2025 also aims to establish a "National System for the Development, Regulation, and Governance of Artificial Intelligence", which would involve public authorities as well as civil society, the private sector and the academia. Inter-ministerial Coordination: The use of the Comitê Interministerial para a Transformação Digital (CITDigital) to ensure a unified national voice and coordinate the implementation of AI solutions across the government. 2. Digital Public Infrastructures (DPIs) and Platforms GOV.BR: A unified digital service platform that centralizes hundreds of public services, utilizing secure authentication and interoperability to deliver high-value solutions to citizens. PIX: An interoperable payment system that serves as a benchmark for how integrated data can foster financial inclusion and digital efficiency. 3. Strategic and Ethical Approaches "Data for Development" (D4D): A paradigm that treats data as a strategic collective asset for the public good and the achievement of Sustainable Development Goals (SDGs), rather than just a commercial commodity Digital Sovereignty: The assertion of the "right to regulate," where states maintain the prerogative to enact laws that protect their citizens, promote local talent, and enhance digital infrastructure without foreign interference Fair Remuneration: A proposal to ensure just and transparent compensation for data providers (citizens, universities, and governments), correcting the "extractive" model where transnational platforms capture all economic value Open-Source and Foundation Models: Prioritizing open-source systems and smaller, specialized "foundation" models to reduce dependence on proprietary closed systems and allow for independent auditing and transparency.