Serbia - AI Ethical Framework

Overview

The Belgrade Ministerial Declaration on Artificial Intelligence represents a landmark international accord, officially adopted by 44 member states of the Global Partnership on Artificial Intelligence (GPAI) and the European Union. This pivotal document emerged from the GPAI Summit held in Belgrade, Serbia, from December 3-4, 2024, with its formal adoption announced on December 27, 2024. Proposed by Serbia, the Declaration signifies a concerted global effort to navigate the transformative potential of AI while ensuring its development and application align with ethical principles and human-centric values. It underscores a collective commitment to fostering an integrated partnership among GPAI members and with the Organisation for Economic Co-operation and Development (OECD), aiming to build a sustainable and trustworthy AI ecosystem that drives economic and social progress while effectively mitigating associated risks and ensuring benefits for all people and the planet.

The Declaration is a testament to Serbia's growing role in global AI discussions, solidifying its position as a leader in advocating for responsible AI governance. It highlights the importance of preserving fundamental values such as individual freedoms, democratic principles, the rule of law, and human rights as the bedrock for all AI development and deployment. Furthermore, the document places a strong emphasis on global cooperation, recognizing that the challenges and opportunities presented by AI transcend national borders. It calls for broad participation from governments, academia, industry, and civil society to collectively work towards an AI future that is inclusive, equitable, and beneficial for humanity, actively seeking to narrow the digital divide between technologically advanced and developing nations.

Definitions

While the Belgrade Ministerial Declaration on AI does not contain a dedicated glossary section, several key concepts are central to its underlying philosophy and intent. 'Human-Centric AI' is a core tenet, referring to the development and deployment of AI systems that prioritize human well-being, rights, and values, ensuring that technology serves humanity rather than dominating it. This involves designing AI systems that are transparent, fair, accountable, and enhance human capabilities, rather than replacing them without due consideration for societal impact. The Declaration implicitly defines 'Trustworthy AI' as AI systems that are reliable, secure, robust, and operate in a manner that inspires confidence among users and the broader public, built upon a foundation of ethical principles and robust governance frameworks.

The Declaration also addresses the concept of 'Ethical AI,' which encompasses the moral principles and values that should guide the entire AI lifecycle, from design and development to deployment and use. These principles include fairness, non-discrimination, privacy, accountability, and the prevention of harm, all aimed at ensuring AI systems contribute positively to society. Furthermore, the document implicitly defines the 'Digital Divide' as the gap in technological advancement and access to digital resources, particularly AI, between different nations and communities. A significant goal of the Declaration is to actively work towards reducing this gap, promoting equitable access to AI technologies and ensuring that all countries, regardless of their current level of technological development, can participate in and benefit from the AI future. This involves fostering inclusive innovation and knowledge sharing to prevent further exacerbation of existing inequalities.

Governance and Institutional Framework

The Belgrade Ministerial Declaration on AI reinforces the Global Partnership on Artificial Intelligence (GPAI) as a crucial multi-stakeholder initiative for advancing responsible AI. It explicitly acknowledges GPAI's foundational achievements and its integrated partnership with the Organisation for Economic Co-operation and Development (OECD), which brings together OECD members and GPAI countries on equal footing, based on the foundational OECD Recommendation on Artificial Intelligence (OECD AI Principles). This framework positions GPAI as a key forum for identifying best practices, conducting technical analysis, and developing international, evidence-based, and consensus-based AI policy. The Declaration highlights the importance of this integrated approach to fostering a sustainable and trustworthy AI ecosystem globally.

Serbia's role as co-chair of GPAI for 2025, alongside the Slovak Republic, is a central element of the Declaration's institutional framework, signaling a commitment to continued leadership in global AI governance discussions. The Declaration calls upon all stakeholders—including governments, policymakers, international and regional organizations, private enterprises of all sizes, academic institutions, technical communities, trade and labor unions, and civil society—to actively engage in advancing its vision. This multi-stakeholder approach is fundamental to ensuring that AI development and regulation are comprehensive, inclusive, and reflective of diverse perspectives and expertise. The document emphasizes that this collaborative effort is essential for developing human-centric, ethical, and rights-respecting AI technologies throughout the entire AI system lifecycle.

Key Focus Areas

The Belgrade Ministerial Declaration on AI articulates several key priorities for the future development and governance of artificial intelligence, placing a strong emphasis on foundational values. A primary focus is the preservation of fundamental values, explicitly reaffirming the importance of individual freedoms, democratic values, the rule of law, and human rights as the essential foundation for all AI development and application. Ministers committed to ensuring that AI technologies respect these values, thereby fostering public trust in new technologies. This commitment aims to prevent AI from undermining societal norms and individual liberties, instead advocating for its use as a tool to uphold and strengthen these principles.

Another critical area of focus is the promotion of global cooperation and the strengthening of international collaboration. The Declaration specifically highlights an integrated partnership with the OECD to collectively build a sustainable and trustworthy AI ecosystem. This collaborative spirit extends to addressing global challenges, with GPAI prioritizing evidence-based research in critical sectors such as employment and labor, data governance and privacy, intellectual property, risk identification, and environmental impacts. The document also acknowledges the transformative potential of AI to help solve the world's most pressing challenges and foster sustainable development, while simultaneously emphasizing the imperative to mitigate associated risks and bridge AI-related digital divides between nations and communities to ensure equitable benefits for all.

Implementation Framework

The implementation framework envisioned by the Belgrade Ministerial Declaration on AI is rooted in sustained international cooperation, multi-stakeholder engagement, and the active integration of scientific expertise. The Declaration commits to building upon the established foundation of the OECD AI Principles and the ongoing work of GPAI. This involves continuous coordination with stakeholders across all sectors and continents to support and promote the safe, secure, and trustworthy development and use of AI technologies. The emphasis is on fostering AI systems that are human-centric, ethical, rights-respecting, and inclusive throughout their entire lifecycle, ensuring that implementation efforts are guided by these core values.

A significant aspect of the implementation strategy involves addressing the global digital divide. The Declaration explicitly aims to reduce the gap in technological development between less and more advanced countries, ensuring that all nations can progress together into the technological future at an equitable pace. This objective is to be achieved through the active involvement of science in all processes related to AI development and deployment, fostering knowledge sharing and capacity building. Furthermore, the Declaration calls upon all stakeholders—governments, international organizations, private sector, academia, and civil society—to engage constructively in advancing its vision, embracing AI as a catalyst for economic and social development while diligently mitigating its associated risks and ensuring benefits for both people and the planet. The ongoing dialogue through GPAI is affirmed as crucial for achieving this shared vision, with science playing a critical role in responding to complex challenges and promoting inclusive development.

Monitoring and Evaluation

While the Belgrade Ministerial Declaration on AI does not prescribe a specific, detailed monitoring and evaluation framework, its spirit and stated objectives inherently call for ongoing assessment of progress and impact. The Declaration underscores the importance of strengthening international collaboration through an integrated partnership with the OECD to build a sustainable and trustworthy AI ecosystem. This partnership, alongside the GPAI's expert community and expert support centers, serves as an implicit mechanism for monitoring the global landscape of AI development and governance. The continuous engagement in constructive dialogue through GPAI is designed to facilitate the exchange of information and best practices, which are essential components of an informal evaluation process regarding the adherence to the Declaration's principles and the effectiveness of implemented strategies.

The commitment to sharing information and GPAI's results with the global community, ensuring that all nations can benefit from advancements in AI, also implies a form of collective evaluation. By promoting transparency and open dialogue, the Declaration encourages a self-correcting mechanism within the international AI community, allowing for the identification of emerging challenges and the adaptation of approaches to foster human-centric and ethical AI. The ongoing work within GPAI, including its focus on evidence-based research in critical sectors, will contribute to understanding the societal and economic impacts of AI, thereby providing continuous feedback that can inform policy adjustments and evaluate the effectiveness of efforts to bridge the digital divide and mitigate risks. This collaborative, iterative process of learning and adaptation forms the de facto monitoring and evaluation approach for this high-level declaration.

Penalties, Liability, and Appeals

As a ministerial declaration, the Belgrade Ministerial Declaration on AI is a high-level political statement and a framework for international cooperation, rather than a legally binding instrument that prescribes specific penalties, liability regimes, or appeal processes. Its primary purpose is to establish shared principles, values, and a common vision for the responsible development and governance of AI. Therefore, the Declaration itself does not contain provisions for fines, sanctions, criminal penalties, or mechanisms for legal liability and redress related to AI systems. These aspects typically fall within the scope of national legislation or more specific international conventions or regulations that are yet to be developed or adopted based on such foundational declarations.

The Declaration does, however, lay the groundwork for future legal and ethical regulation of AI by emphasizing the need for mechanisms to legally and ethically regulate AI. It calls upon governments and policymakers to engage in advancing this vision, which implicitly includes the eventual development of robust legal frameworks that would address issues of liability, enforcement, and redress. While the Declaration itself does not impose direct legal obligations or penalties, its adoption by numerous states and the European Union signals a collective intent to move towards a regulatory environment where such provisions would be established. The emphasis on human rights, the rule of law, and risk mitigation within the Declaration suggests that future legal instruments developed in its spirit would indeed incorporate strong accountability measures to protect individuals and ensure responsible AI deployment.

Relationship to Other Instruments

The Belgrade Ministerial Declaration on AI explicitly positions itself within a broader landscape of international AI governance instruments, building upon and reinforcing existing frameworks. It notably recalls the ministerial outcomes of previous GPAI summits, most recently the 2024 New Delhi GPAI Declaration. This earlier declaration announced a renewed vision for GPAI through an integrated partnership with the OECD, bringing together all current OECD members and GPAI countries on equal footing. This continuity highlights a consistent trajectory in global efforts to establish a unified approach to AI governance.

Crucially, the Belgrade Declaration confirms its foundation on the OECD Recommendation on Artificial Intelligence (OECD AI Principles). These principles are recognized for their importance in promoting the use of AI that is innovative, trustworthy, and respectful of human rights and democratic values. By explicitly committing to build on these principles, the Belgrade Declaration reinforces their status as a cornerstone of international AI policy. This synergistic relationship ensures that the Declaration's objectives are consistent with widely accepted international standards and best practices, fostering a coherent and harmonized approach to AI regulation across diverse jurisdictions and promoting global interoperability in AI governance frameworks.

International Alignment

International alignment is a cornerstone of the Belgrade Ministerial Declaration on AI, which explicitly aims to strengthen global cooperation and foster an integrated partnership among GPAI members and with the OECD. The Declaration underscores the shared commitment to building a sustainable and trustworthy AI ecosystem, recognizing that the challenges and opportunities presented by AI are inherently global. It reaffirms the importance of the OECD AI Principles as a common foundation for promoting innovative, trustworthy, and human rights-respecting AI, thereby seeking to harmonize approaches to AI governance across participating nations.

The Declaration also expresses a forward-looking commitment to welcoming new members, particularly developing and emerging economies, to GPAI. This inclusive approach aims to reinforce collaboration with non-members and other international organizations, ensuring that a broad range of expertise and national and regional perspectives are represented in global AI discussions. By promoting broad global participation and sharing information and results from GPAI's work, the Declaration seeks to ensure that all nations can benefit from advancements in AI, actively working to bridge the digital divide and foster equitable technological development worldwide. This emphasis on comprehensive international alignment is critical for establishing a unified and effective global framework for responsible AI.

Implementation Timeline

Compliance Checklist

Sources and References