Consulta Pública de la Estrategia Nacional de Inteligencia Artificial
AGESIC
Submission of The Center for AI and Digital Policy (CAIDP) to the Agency for Development of Electronic Government and Information Society (AGESIC) of
Aceptadas
The Center for AI and Digital Policy (CAIDP) appreciates the opportunity to provide comments to the Uruguay Artificial Intelligence (AI) Strategy 2024 - 2030 Public Consultation. We commend Uruguay's efforts in developing a forward-looking AI strategy that aligns with international and regional frameworks on ethical AI, including the UNESCO Recommendation on the Ethics of AI that Uruguay had endorsed in 2021 , and the recently adopted Declaration of Montevideo For the construction of a regional approach on the governance of AI and its impacts on our society.
About CAIDP
The Center for AI and Digital Policy (CAIDP) is a global, independent, non-profit research organization headquartered in Washington, D.C. CAIDP’s mission is to promote a more just society by ensuring that technology advances uphold human rights, democratic governance, and the rule of law. CAIDP offers expert guidance on AI policy to a variety of international organizations, including UNESCO, the OECD, the G20, the European Parliament, and others. Additionally, CAIDP publishes the annual Artificial Intelligence and Democratic Values (AIDV) Report, which provides a comprehensive evaluation of global AI policies and practices, using a standardized methodology and twelve specific metrics to assess alignment with democratic principles. The most recent edition of the AIDV Report (2023) highlighted that Uruguay did not yet have a general national AI development strategy, but rather has an AI development strategy for the digital government, officially called Artificial Intelligence Strategy for Digital Government (AISDG) 2020. CAIDP notes Uruguay’s efforts towards formulating a general AI strategy with the draft National AI Strategy 2024 – 2030 that is open for public consultation.
CAIDP Comments on Draft Uruguay AI Strategy’s Alignment with Regional and International Ethical AI Frameworks
CAIDP notes that the draft Uruguay AI Strategy demonstrates alignment with regional and international ethical AI frameworks. The Uruguay AI Strategy embodies key elements of the Montevideo Declaration, particularly its focus on human rights, transparency, and multi-stakeholder governance. Specifically, the Uruguay AI Strategy's emphasis on human rights (as seen in its principles of respect for human dignity and non-discrimination, detailed in Section 'Guiding Principles') aligns with the Montevideo Declaration's call for AI development that respects and promotes fundamental freedoms and human rights. The draft Uruguay AI Strategy also incorporates transparency and accountability mechanisms, such as the establishment of the AI Observatory (detailed in Section 'Axis 1: Governance'), which includes monitoring AI initiatives, evaluating their societal impact, and ensuring compliance with ethical standards. The proposed AI Observatory will include representatives from government bodies, academia, civil society organizations, and the private sector, ensuring a comprehensive multi-stakeholder approach (as detailed in Section 'Axis 1: Governance') , which aligns with the Montevideo Declaration's emphasis on regional AI governance that includes multiple stakeholders and ensures transparency.
The draft Uruguay AI Strategy also reflects the principles of sustainable technological development and inclusivity, as emphasized in the Montevideo Declaration. For example, Axis 3 of the draft Uruguay AI Strategy, which focuses on sustainable development and the sectors of the national economy , directly supports the Montevideo Declaration's goals of ensuring that AI promotes economic development in a sustainable and inclusive manner.
Furthermore, the guiding principles articulated in the draft Uruguay AI Strategy (detailed in Section 'Guiding Principles') are consistent with the UNESCO Recommendation on the Ethics of AI, including respect for human dignity, equitable access, and the importance of education and training for AI capabilities. The draft Uruguay AI Strategy’s guiding principles, such as those focusing on equity, non-discrimination, and strengthening democracy, directly align with the UNESCO Recommendation’s emphasis on protecting human rights and fostering inclusiveness (as detailed in Section 'Values and Principles', and 'Policy Area 1: Ethical Impact Assessment'). Additionally, the Strategy’s emphasis on talent management and AI education initiatives (as detailed in Section 'Axis 2: AI Capabilities') aligns with the UNESCO Recommendation’s focus on the importance of education and research to empower societies in the AI era (as detailed in Section 'Policy Area 8: Education and Research').
Furthermore, the guiding principles articulated in the draft Uruguay AI Strategy (detailed in Section 'Guiding Principles') are consistent with the UNESCO Recommendation on the Ethics of AI, including respect for human dignity, equitable access, and the importance of education and training for AI capabilities. The draft Uruguay AI Strategy’s guiding principles, such as those focusing on equity, non-discrimination, and strengthening democracy, directly align with the UNESCO Recommendation’s emphasis on protecting human rights and fostering inclusiveness (as detailed in Section 'Values and Principles', and 'Policy Area 1: Ethical Impact Assessment'). Additionally, the Strategy’s emphasis on talent management and AI education initiatives (as detailed in Section 'Axis 2: AI Capabilities') aligns with the UNESCO Recommendation’s focus on the importance of education and research to empower societies in the AI era (as detailed in Section 'Policy Area 8: Education and Research').
CAIDP Recommendations
CAIDP commends Uruguay for taking steps to ensure that AI technologies contribute to well-being while mitigating risks to human rights and democracy, including the explicit incorporation of mechanisms for monitoring and review, which support both the Montevideo Declaration and UNESCO's recommendations for ongoing ethical AI governance. CAIDP proposes several recommendations that would strengthen the draft Uruguay AI strategy through stronger alignment with the Montevideo Declaration and the UNESCO Recommendation on the Ethics of AI, which have both been endorsed by Uruguay.
1. Adoption of a stronger human rights based approach
We recommend the adoption of a stronger human rights-based approach to guide the development and deployment of AI in Uruguay. This approach should explicitly integrate safeguards that protect vulnerable communities and uphold fundamental rights, which aligns with both the Montevideo Declaration and the UNESCO Recommendation. For example, requirements for ethical impact assessments (as emphasized in Section 'Ethical Impact Assessment' of the UNESCO Recommendation) and consultations with affected communities (as highlighted in the Montevideo Declaration) could further enhance the Strategy's alignment with international best practices already endorsed by Uruguay.
We recommend the adoption of a stronger human rights-based approach to guide the development and deployment of AI in Uruguay. This approach should explicitly integrate safeguards that protect vulnerable communities and uphold fundamental rights, which aligns with both the Montevideo Declaration and the UNESCO Recommendation. For example, requirements for ethical impact assessments (as emphasized in Section 'Ethical Impact Assessment' of the UNESCO Recommendation) and consultations with affected communities (as highlighted in the Montevideo Declaration) could further enhance the Strategy's alignment with international best practices already endorsed by Uruguay.
2. Enhance Explicit Transparency and Accountability Measures
The draft Uruguay AI Strategy outlines commitments to transparency and explainability, which is commendable. CAIDP suggests incorporating mandatory periodic third-party audits for AI systems, particularly those used in public services. Independent audits would help ensure compliance with ethical guidelines and bolster public trust in AI technologies. This recommendation strengthens the Uruguay AI Strategy's commitment to transparency and explainability by adding an external layer of verification that ensures accountability. Periodic third-party audits would provide an objective assessment of AI systems, complementing the draft Uruguay AI Strategy's existing mechanisms such as the AI Observatory, and ensuring that transparency is maintained throughout the entire lifecycle of AI technologies (as detailed in Section 'Policy Area 2: Ethical Governance and Stewardship' of the UNESCO Recommendation).
The draft Uruguay AI Strategy outlines commitments to transparency and explainability, which is commendable. CAIDP suggests incorporating mandatory periodic third-party audits for AI systems, particularly those used in public services. Independent audits would help ensure compliance with ethical guidelines and bolster public trust in AI technologies. This recommendation strengthens the Uruguay AI Strategy's commitment to transparency and explainability by adding an external layer of verification that ensures accountability. Periodic third-party audits would provide an objective assessment of AI systems, complementing the draft Uruguay AI Strategy's existing mechanisms such as the AI Observatory, and ensuring that transparency is maintained throughout the entire lifecycle of AI technologies (as detailed in Section 'Policy Area 2: Ethical Governance and Stewardship' of the UNESCO Recommendation).
3. Require Mandatory Ethical Impact Assessments
CAIDP proposes that the draft Uruguay AI Strategy include a requirement for mandatory ethical impact assessments before the deployment of high-risk AI applications. Such assessments should cover potential impacts on human rights, privacy, and environmental sustainability. This recommendation supports proactive risk management and aligns with UNESCO's emphasis on ethical governance throughout the AI lifecycle (as detailed in Section 'Policy Area 2: Ethical Governance and Stewardship', of the UNESCO Recommendation).
CAIDP proposes that the draft Uruguay AI Strategy include a requirement for mandatory ethical impact assessments before the deployment of high-risk AI applications. Such assessments should cover potential impacts on human rights, privacy, and environmental sustainability. This recommendation supports proactive risk management and aligns with UNESCO's emphasis on ethical governance throughout the AI lifecycle (as detailed in Section 'Policy Area 2: Ethical Governance and Stewardship', of the UNESCO Recommendation).
4. Prohibit High-Risk Applications
CAIDP urges Uruguay to include in the draft AI Strategy explicit prohibitions of certain high-risk AI applications that infringe on fundamental rights, such as mass surveillance and discriminatory profiling. This step would affirm Uruguay's commitment to the ethical use of AI and prevent misuse that could harm marginalized groups, in line with the principles outlined in both the Montevideo Declaration and UNESCO Recommendation (Section 'Values and Principles').
CAIDP urges Uruguay to include in the draft AI Strategy explicit prohibitions of certain high-risk AI applications that infringe on fundamental rights, such as mass surveillance and discriminatory profiling. This step would affirm Uruguay's commitment to the ethical use of AI and prevent misuse that could harm marginalized groups, in line with the principles outlined in both the Montevideo Declaration and UNESCO Recommendation (Section 'Values and Principles').
5. Continuous Inclusive Multi-Stakeholder Engagement
CAIDP recommends that the draft Uruguay AI Strategy establish formal mechanisms for continuous engagement with civil society, academia, and the private sector to ensure diverse perspectives are integrated throughout the policy implementation process. This recommendation builds on and strengthens the AI Observatory proposal in the draft Uruguay AI Strategy by promoting a more structured and continuous form of engagement beyond the initial setup. The AI Observatory, as proposed in the draft Uruguay AI Strategy (Section 'Axis 1: Governance') , already includes representatives from key stakeholder groups. CAIDP’s recommendation aims to ensure that these representatives not only participate but are also actively involved in ongoing decision-making and monitoring processes to adapt to emerging challenges and opportunities in AI governance. This ongoing dialogue is essential to reflect the evolving nature of AI technologies and the ethical considerations they entail.
CAIDP recommends that the draft Uruguay AI Strategy establish formal mechanisms for continuous engagement with civil society, academia, and the private sector to ensure diverse perspectives are integrated throughout the policy implementation process. This recommendation builds on and strengthens the AI Observatory proposal in the draft Uruguay AI Strategy by promoting a more structured and continuous form of engagement beyond the initial setup. The AI Observatory, as proposed in the draft Uruguay AI Strategy (Section 'Axis 1: Governance') , already includes representatives from key stakeholder groups. CAIDP’s recommendation aims to ensure that these representatives not only participate but are also actively involved in ongoing decision-making and monitoring processes to adapt to emerging challenges and opportunities in AI governance. This ongoing dialogue is essential to reflect the evolving nature of AI technologies and the ethical considerations they entail.
The Center for AI and Digital Policy (CAIDP) extends its appreciation to the Agency for Development of Electronic Government and Information Society (AGESIC) of the Oriental Republic of Uruguay for considering our comments and recommendations regarding the Draft Uruguay National Artificial Intelligence Strategy 2024 - 2030. We welcome the opportunity for continued engagement and remain available to address any further inquiries.
Sincerely,
Marc Rotenberg Merve Hickok
CAIDP Executive Director CAIDP President
CAIDP Executive Director CAIDP President
Julian Theseira
Policy Group Leader (Global Majority)
Policy Group Leader (Global Majority)
Esta propuesta ha sido aceptada porque:
Muchas gracias por su participación en la consulta pública.
A continuación, se detallan los elementos del aporte recibido integrados y aquellos que no fueron incorporados.
- Adopción de un enfoque más sólido basado en los derechos humanos. La sugerencia se integró parcialmente. Se incorporó un párrafo al Principio 3, Respeto de la dignidad y los derechos humanos, relativo a los impactos diferenciados implicados en los grupos en situación de vulnerabilidad, y una acción específica en la sección: Eje 3. Desarrollo sostenible. Línea de acción 3.3 Impacto de la IA en la Sociedad. Acción 8. En cuanto a la recomendación asociada a la evaluación de impacto ético, no fue integrada en función, de dos elementos: que la Estrategia ya prevé la generación de instrumentos de evaluación de impacto (Sección Eje 1. Gobernanza. Línea de acción 1.2 Marco regulatorio. Acción 5), y en tanto que, para el análisis de las contribuciones recibidas en el marco de la consulta pública se definió como criterio que todas las soluciones regulatorias específicas deberán ser discutidas y analizadas en la etapa de implementación de la Estrategia, tal y como está previsto en la Línea 1.2 Marco regulatorio, en la que se estableció como objetivo desarrollar y adoptar los marcos regulatorios necesarios. En función de este criterio general, sólo fueron integrados a la versión final de la Estrategia elementos orientados a precisar enfoques regulatorios o a especificar las temáticas a abordar en la discusión regulatoria posterior, descartándose en esta etapa aquellas que ingresaban en soluciones específicas.
- Mejorar las medidas explícitas de transparencia y rendición de cuentas. La recomendación no fue incorporada en función del criterio relacionado con el contenido específico del marco regulatorio que fue explicitado en el punto anterior.
- Exigir evaluaciones de impacto ético obligatorias. La recomendación no fue incorporada en función del criterio explicitado en el primer punto.
- Prohibir aplicaciones de alto riesgo. La recomendación fue integrada parcialmente a la Estrategia en función del criterio mencionado anteriormente. En tal sentido se entendió que explicitar la necesidad de consensuar las aplicaciones de los sistemas de IA que resultan inaceptables como sociedad por su impacto sobre los derechos de las personas, evaluando las prohibiciones o moratorias que resulten adecuadas, resultaba pertinente en tanto supone “especificar las temáticas a abordar en la discusión regulatoria” posterior. Esta acción es consistente con lo planteado por la Agesic en el Informe artículo 74 Ley N° 20.212 Recomendaciones para una regulación de la Inteligencia Artificial (IA) orientada al desarrollo ético, la protección de los derechos humanos y el fomento de la innovación tecnológica.
- Participación continua e inclusiva de múltiples partes interesadas. La recomendación fue integrada a la Estrategia en la sección: Eje 1. Gobernanza. Línea de acción 1.1 Institucionalidad. Acción 1.
Compartir: