Ministry of Foreign Affairs of Peru

A successful first Global Dialogue on AI Governance should deliver practical, balanced and development-oriented outcomes that help countries move from broad principles to implementation. For Peru, success would begin with consolidating the Dialogue as an inclusive intergovernmental platform that reduces fragmentation in global AI discussions and promotes greater coherence across existing initiatives. A first important outcome would be a clear focus on capacity-building for developing countries, especially in digital infrastructure, access to computing power, data governance, public-sector readiness and specialized talent. These elements remain essential for meaningful participation in the AI ecosystem. A second desirable outcome would be the exchange of concrete national experiences and usable governance tools, including risk assessments, regulatory sandboxes, coordination mechanisms and public-sector guidance. This would help countries identify what works in practice and adapt successful approaches to their own contexts. A third outcome should be stronger convergence around safe, secure and trustworthy AI, grounded in transparency, accountability and meaningful human oversight, while also recognizing the importance of interoperable data systems, trusted data-sharing arrangements and internationally recognized standards. Ultimately, the Dialogue will be successful if it identifies concrete avenues for cooperation, shared learning and measurable follow-up, particularly for countries still building the foundations for effective AI governance. In that sense, it should strengthen implementation capacity, reduce governance gaps and support more inclusive participation in shaping global AI governance.

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

Peru notes that the thematic areas reflected in this questionnaire do not fully correspond to the structure and wording of General Assembly resolution 79/325. The present selection therefore seeks to remain substantively aligned with the resolution while adapting to the format of the questionnaire. Capacity-building remains a central priority. For developing countries, meaningful participation in the AI ecosystem depends on stronger digital foundations, including computing infrastructure, data governance, specialized skills and institutional preparedness. The social, economic, ethical, cultural, linguistic and technical implications of artificial intelligence are also of particular importance. Peru considers that AI governance should reflect the broader development context in which these technologies are deployed, including inclusion, linguistic diversity and equitable access to their benefits. Interoperability and compatibility of governance approaches is another priority. Peru supports greater compatibility across national and regional approaches, based on cooperation, practical exchange and common reference points, while avoiding the imposition of a single regulatory model. Finally, transparency, accountability and robust human oversight, in a manner consistent with international law, are indispensable to fostering trust in AI systems and ensuring that their development and use remain aligned with public interest objectives and human rights considerations.

Yes. While the listed themes capture many core dimensions of AI governance, Peru considers that several cross-cutting and emerging issues deserve more explicit attention. First, data governance should be recognized more clearly as an enabling foundation for trustworthy AI. The availability, quality, interoperability and security of public datasets are essential for reliable AI systems, especially in the public sector. Peru's current experience shows that weak data standardization, low publication and use of datasets, and uneven adoption of interoperability tools remain significant barriers to effective implementation. Second, institutional capacity and governance architecture deserve more explicit attention. Effective AI governance depends not only on principles and safeguards, but also on having institutions with the authority, coordination capacity and technical means to implement, supervise and scale responsible AI policies. In Peru, this institutional strengthening dimension is increasingly relevant in light of ongoing efforts to reinforce national digital governance. Third, greater visibility should be given to computing infrastructure and equitable access to advanced computational resources. For many developing countries, the AI divide is also about access to high-performance computing, cloud services, digital infrastructure and technical capacity to deploy systems at scale. Peru's ENIA identifies these as persistent constraints. Fourth, Peru sees value in explicitly recognizing implementation and public-sector readiness as a cross-cutting issue. Beyond high-level principles, countries need practical tools, institutional coordination, guidance for public administration and mechanisms to scale successful use cases responsibly. Finally, Peru considers it important to preserve a development-oriented approach, aligned with internationally recognized frameworks, including relevant OECD standards, while taking into account different national capacities and development needs.

Governance gaps in Peru are affecting the country's ability to deploy AI in a strategic, coordinated and trustworthy manner. The most significant challenge is institutional fragmentation. AI adoption across the public sector remains uneven, with different levels of maturity, weak coordination, limited data governance, and major gaps in computing infrastructure and specialized talent. These factors reduce the State's ability to move from reactive service delivery to more predictive, efficient and citizen-centered public services. A second challenge is limited implementation capacity and duplication of efforts. Recent national analysis has shown fragmented digital investments, duplication in infrastructure and software projects, weak interoperability, and disjointed information systems across public entities. This weakens efficiency, slows innovation, and limits the scalability of digital and AI solutions. A third challenge concerns trustworthy governance. Risks related to transparency, accountability, human oversight, cybersecurity and public trust continue to shape how AI is perceived and adopted. If not properly addressed, these gaps may reduce trust and slow responsible use of AI in both public and private sectors. At the same time, Peru has important opportunities. The country already has a legal and regulatory framework, including Law No. 31814, its Regulation, and ISO/IEC-based standards. It also has emerging public-sector use cases, growing private-sector interest, and a proposed institutional strengthening process through the National Agency for Digital Transformation (ANTD) and the CNIDIA. Together, these developments create an opportunity to strengthen AI governance in a way that is human-centered, development-oriented and aligned with international standards, including those of the OECD.

The AI Dialogue can play a valuable role as an inclusive, practical and cooperation-oriented platform that helps translate broad principles into more coordinated international action. First, it can help reduce fragmentation across the growing number of AI-related initiatives by promoting exchange among Member States and other relevant stakeholders, while identifying areas of convergence across governance approaches. For countries such as Peru, this is important to avoid a fragmented landscape in which regulatory and policy discussions advance without sufficient coherence or consideration of development needs. Second, the Dialogue can strengthen capacity-building and international support for developing countries. In practice, cooperation on AI governance must go beyond norms and include access to infrastructure, computing resources, data governance tools, institutional training and technical assistance. The Dialogue can help identify these needs and connect them with concrete avenues for cooperation. Third, it can serve as a platform for sharing practical national experiences, including governance tools, public-sector applications, regulatory approaches and risk-management practices. This type of exchange can help countries learn from one another and adapt successful approaches to their own contexts. Fourth, the Dialogue can promote greater understanding of the enabling role of interoperable data systems, trusted data-sharing arrangements and technical standards in supporting safe, secure and trustworthy AI. Ultimately, the AI Dialogue should help advance a form of international cooperation that ensures developing countries participate meaningfully in shaping global AI governance and in benefiting from its opportunities.

The AI Dialogue should build upon and connect with existing United Nations mechanisms, particularly the Independent International Scientific Panel on AI, so that scientific evidence can better inform intergovernmental deliberations and help avoid duplication between technical assessment and policy discussion. Peru has already highlighted the importance of linking the Dialogue and the Panel in a structured and policy-oriented manner. It should also connect with regional and interregional processes, including those involving ECLAC/CEPAL in Latin America, so that global discussions are informed by regional realities, development needs and practical experiences. This would help translate broad global discussions into more actionable cooperation agendas. In addition, the Dialogue should build on existing international principles, standards and governance approaches. In general terms, Peru considers that international cooperation on AI governance should help countries progressively align with internationally recognized standards, particularly relevant OECD standards and other risk-based frameworks, while taking into account different national capacities and development needs. The Dialogue should also connect with national institutional processes aimed at strengthening AI and digital governance. In Peru, this includes the proposed National Agency for Digital Transformation (ANTD), which is intended to strengthen governance, coordination, supervision and implementation across digital transformation, data governance, digital security and artificial intelligence. Its added value should be to serve as a practical intergovernmental platform that reduces fragmentation, promotes coherence across initiatives, and translates broad principles into concrete cooperation-oriented outcomes, especially for developing countries.

Different stakeholders can contribute to the AI Dialogue through structured, sector-based and multi-stakeholder mechanisms that combine policy leadership, technical expertise and implementation experience. In Peru's experience, one useful model has been the organization of thematic working arrangements under the national digital agenda. For example, the Technical Working Group on the Digital Economy is led by the Ministry of Production, through the Directorate of Digitalization and Formalization, and is tasked with contributing to the diagnosis and identification of gaps in the field of the digital economy. At the broader level, the PCM established the Multisectoral Working Group to develop the Digital Agenda of Peru to 2030, which reflects the wider coordinating role of the Secretariat of Government and Digital Transformation (SGTD) as the governing body of the National Digital Transformation System. In addition, Peru has used the Technical Working Group of the Peruvian Digital Agenda entrusted to the Ministry of Foreign Affairs to follow up on international commitments and external engagement. These spaces bring together not only public institutions, but also private sector representatives, representatives of foreign States, and international organizations, allowing for broader, informed and coordinated contributions to public policy discussions. Based on this experience, Member States should remain at the center of the AI Dialogue, while other stakeholders contribute through clearly defined channels. International organizations can provide comparative knowledge and technical assistance; the private sector can share innovation and implementation experience; academia and technical experts can contribute research and evidence; and civil society can help reflect social impacts and inclusion concerns. As for format and structure, the Dialogue would benefit from a modular and results-oriented design, with thematic tracks, structured working groups, and dedicated stakeholder segments linked to specific policy areas. This should be supported by a clear roadmap, balanced participation, and concrete outputs, so that discussions move beyond general debate and contribute to practical cooperation and implementation.

Global discussions on AI governance still tend to underrepresent the perspectives of developing countries, particularly those with more limited technological, financial and institutional capacities. Their participation is essential because they face some of the most significant barriers to AI readiness, including gaps in computing infrastructure, access to data, specialized talent and public-sector capacity. Without their meaningful inclusion, global governance frameworks risk being shaped primarily by the realities of more advanced economies. There is also insufficient representation of public-sector practitioners, especially from countries that are still building institutional frameworks for AI governance. Their experience is valuable because many implementation challenges arise not only at the normative level, but also in service delivery, procurement, coordination, oversight and the responsible use of data in government. In addition, greater space should be given to regional perspectives, including those of Latin America and other regions whose development priorities, linguistic diversity and institutional conditions are not always adequately reflected in global debates. Peru has emphasized that governance discussions should remain attentive to inclusion, development needs and diverse national trajectories. These gaps could be addressed through a more balanced and structured participation model. This includes financial and logistical support for developing countries, geographically balanced facilitation, stronger links with regional processes, and dedicated spaces for practical contributions from public institutions, academia, civil society and the private sector, with clear safeguards to avoid disproportionate influence or conflicts of interest. It is also important that contributions from non-State actors complement, rather than displace, the central role of Member States in the Dialogue.

Innovative engagement formats should help make the AI Dialogue more practical, inclusive and implementation-oriented, rather than limited to general statements. One useful format would be thematic modular sessions with a limited number of standing tracks and continuity across annual cycles. This would allow participants to move beyond one-off discussions and build cumulative work on priority issues such as capacity-building, interoperability and trustworthy AI. A second format could be policy labs or implementation roundtables, where Member States and relevant stakeholders present concrete national experiences, regulatory tools, pilot initiatives or public-sector use cases, followed by structured discussion on lessons learned, scalability and cooperation needs. This would be particularly valuable for countries seeking practical guidance rather than abstract principles alone. A third option would be science-policy interface sessions involving the Independent International Scientific Panel on AI, with findings presented in concise, policymaker-oriented formats and followed by interactive exchanges with delegations. Peru has already highlighted the value of a structured mechanism to translate scientific expertise into decision-relevant discussions. The Dialogue could also benefit from regional breakout sessions or cross-regional exchanges, allowing participants to reflect specific development realities, regional priorities and cooperation opportunities, and then bring those insights back into the global discussion. Finally, dynamic engagement would be strengthened through multi-stakeholder segments with clear rules, where academia, civil society, the private sector and technical experts provide targeted inputs in designated spaces, while preserving the intergovernmental character of the process and ensuring transparency and balanced participation.

Examples from Peru that promote effective AI governance and offer concrete solutions include: 1. Law No. 31814 and its Regulation (Supreme Decree No. 115-2025-PCM). These establish a formal framework for the responsible use, development and promotion of AI, while assigning the SGTD a central coordination and oversight role. This helps address fragmentation and clarify institutional responsibilities. 2. Mandatory adoption of NTP-ISO/IEC 42001:2025 by public entities implementing AI systems. This introduces a practical risk-based management approach, including transparency, auditability, continuous improvement and human oversight. 3. Data governance and interoperability tools. Peru already has the National Open Data Platform, the National Interoperability Policy, and the State Interoperability Platform (PIDE). These mechanisms help improve data availability, exchange and standardization, which are essential for reliable and scalable AI. 4. The proposed National Center for Digital Innovation and Artificial Intelligence (CNIDIA). This is intended to serve as a national hub for high-performance computing, experimentation, research and innovation, helping address current gaps in computing infrastructure and technical capacity. 5. Public-sector AI applications already in use. Examples include CURIA in the Judiciary, ELECCIA in the electoral system, and RENATA in civil registry services. These cases show how AI can improve efficiency, service delivery and traceability when supported by governance safeguards. 6. Regulatory sandboxes and impact assessments. Peru identifies these as useful tools to test AI solutions in controlled environments and manage high-risk uses more responsibly. Taken together, these examples show that effective AI governance requires standards, institutions, infrastructure and practical implementation tools.