Semantix

The success of the first Global Dialogue on AI Governance should be measured by its ability to promote inclusive, decentralized governance oriented toward the common good. One of the key expected outcomes is the decentralization of technological power, respecting the sovereignty of each country. AI governance cannot remain concentrated in a few nations or large corporations. It must be built collaboratively, taking into account diverse local, cultural, and economic contexts, ensuring that all countries have an active voice in shaping guidelines and standards. Another central point is addressing self-regulation driven solely by the interests of large corporations. When unbalanced, this model can deepen power asymmetries and exacerbate inequalities. Therefore, the dialogue must advance the development of transparent, accountable, and multi-stakeholder governance mechanisms, involving governments, civil society, academia, and the private sector. Furthermore, it is essential to establish concrete commitments to international cooperation, with a focus on knowledge transfer, technology sharing, and capacity building for developing countries. This will enable AI to serve as a tool for inclusion, fostering sustainable development and reducing global inequalities rather than amplifying them. Finally, true success will lie in the ability to translate principles into practical actions, with clear, measurable goals that are monitored over time. Diversity, equity, and collaboration must move beyond aspirational values and become operational pillars of global AI governance.

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

My choices reflect the need to build a global AI governance framework that is decentralized, inclusive, and oriented toward reducing inequalities. Prioritizing safe, secure, and trustworthy AI is essential; however, this security must also include addressing algorithmic bias and injustice. AI systems, when trained on unbalanced or non-diverse data, can reproduce and even amplify social, economic, racial, and cultural discrimination. Therefore, ensuring algorithmic fairness is critical to prevent AI from reinforcing existing inequalities. The development of AI capabilities is another central pillar, especially for developing countries. Without access to knowledge, infrastructure, and education, these countries remain excluded from both the creation and governance of these technologies. Promoting capacity building is a key step toward democratizing access and enabling more equitable and active participation in the global landscape. Furthermore, considering the social, economic, ethical, cultural, linguistic, and technical implications of AI expands governance beyond a purely technological perspective. This approach allows for the inclusion of diverse realities and contexts, ensuring that AI is developed and deployed responsibly, with respect for diversity and human rights. These priorities reinforce the importance of collaborative and decentralized governance, in which major global actors also take responsibility for reducing inequalities, supporting sustainable development, and promoting social justice through AI.

Yes, there are cross-cutting and emerging issues that still require greater attention in the global debate on AI governance. One of them is the concentration of technological and economic power, which remains a structural risk. Despite advances in governance, a small number of countries and large corporations still dominate infrastructure, data, and computational capacity, potentially limiting the digital sovereignty of other nations. Another critical point is data governance, particularly regarding the origin, quality, and representativeness of the data used to train AI models. The lack of diversity in datasets is directly linked to algorithmic bias and injustice, reinforcing historical inequalities and disproportionately affecting underrepresented groups. It is also important to highlight the need for global accountability mechanisms. There are still significant gaps in determining responsibility for decisions made by AI systems, especially in high-impact areas such as healthcare, credit, and public safety. In addition, the sustainability of AI is an emerging concern. The high energy consumption and environmental impact of large-scale models must be considered within governance strategies. Finally, the meaningful inclusion of the Global South in decision-making processes remains a challenge. Participation must go beyond symbolic representation and ensure real influence in shaping global policies, standards, and frameworks. These issues reinforce the need for a more integrated, ethical, and collaborative approach that goes beyond traditional themes and addresses the structural challenges of AI governance.

In the context of Brazil and the Global South, gaps in AI governance reveal structural challenges that directly impact vulnerable populations. A critical issue is that data protection laws, while essential, are still insufficient to effectively protect the most vulnerable. Many of these regulations are inspired by Global North models and reflect a Eurocentric perspective, which does not always align with the social, economic, and cultural realities of the Global South. As a result, there is a misalignment between regulation and the actual needs of the population. This becomes even more concerning in the use of data for AI systems. When legal frameworks fail to account for structural inequalities, the use of data can reinforce exclusion rather than mitigate it. Populations with less access to rights, infrastructure, and representation are often the least protected in these systems, increasing the risks of discrimination and algorithmic invisibility. Furthermore, dependence on external technologies and standards limits the ability of Brazil and other Global South countries to develop solutions aligned with their own priorities. This directly impacts digital sovereignty and the development of more just and context-aware governance. However, this scenario also presents a strategic opportunity: to rethink governance models based on local realities, embedding diversity, equity, and inclusion as core principles rather than secondary considerations. Addressing these gaps requires not only regulatory adaptation but a transformation in how AI governance is conceived, ensuring it truly protects those who need it most.

The Global Dialogue on AI can play a decisive role in transforming governance principles into concrete mechanisms for implementation, auditing, and accountability. One of its main roles is to move beyond conceptual frameworks and advance toward the definition of common operational standards. This includes the creation of clear metrics, risk indicators, and practical guidelines that enable countries to implement AI governance in a consistent and measurable way. In addition, the dialogue can drive the development of international auditing and accountability mechanisms. This involves establishing independent evaluation structures capable of monitoring AI systems, identifying biases, risks, and impacts, and ensuring compliance with ethical and regulatory principles. Without auditing, governance remains merely declarative. Another essential role is to foster technical cooperation among countries, particularly by supporting the Global South through capacity building, knowledge transfer, and infrastructure development. This ensures that governance is not only normative but also practically feasible. The dialogue can also encourage the standardization of practices such as algorithmic impact assessments, model documentation, and data transparency, creating a common foundation across countries and sectors. Finally, it is crucial to establish binding commitments or continuous monitoring mechanisms to ensure that agreements go beyond intent and are effectively implemented over time. In this way, the AI Dialogue can evolve from a discussion platform into a driver of global execution, promoting governance that is effective, auditable, and results-oriented.

The AI Dialogue should build on existing initiatives to avoid fragmentation and accelerate more coherent global governance. Key references include the principles of the OECD, the AI Ethics Recommendation by UNESCO, efforts by the United Nations, as well as initiatives such as the Global Partnership on AI and the EU AI Act. Despite these efforts, fragmentation and implementation gaps remain, particularly in the Global South. The value of the Dialogue lies in acting as a global connector—aligning frameworks and translating principles into practical action. It can also amplify the voice of the Global South, ensure local adaptation, and strengthen international cooperation. Additionally, it can promote shared standards, auditing mechanisms, and common metrics, enhancing trust and the overall effectiveness of AI governance.

Different stakeholders play a key role in making the AI Dialogue inclusive and effective. Governments should lead regulation and ensure accountability; the private sector should contribute with transparency and responsible innovation; academia provides rigor and evaluation; and civil society ensures representation of vulnerable groups. The Dialogue should adopt an action-oriented structure, combining strategic discussions with technical working groups focused on implementation (such as auditing and risk assessment). Regional representation, especially from the Global South, is essential to avoid one-size-fits-all approaches. It should also be continuous, with implementation and monitoring cycles, clear goals, and knowledge sharing. Co-creation spaces across sectors are critical to turning dialogue into concrete impact.

Voices from the Global South, Indigenous communities, Black and marginalized populations, women, and workers impacted by AI remain underrepresented in global discussions. These groups are often the most affected but have limited influence, while decisions are largely shaped by Global North perspectives. To include them effectively, it is necessary to ensure meaningful participation with real decision-making power, invest in capacity building and access to technology, and adopt participatory approaches such as public consultations and co-creation. Promoting diversity in teams and datasets is also essential to reduce bias. Inclusion must be treated as a structural component of AI governance to ensure more just and representative outcomes.

To promote more meaningful and dynamic engagement in the AI Dialogue, it is essential to leverage AI itself as a tool for inclusion and the democratization of voices. AI can play a central role by enabling real-time multilingual participation through automatic translation and language accessibility, allowing people from different regions to actively engage without language barriers. This is key to ensuring more diverse and global representation. In addition, AI-powered digital platforms can expand the reach of the dialogue by enabling asynchronous and continuous contributions, particularly from communities that have been historically excluded from decision-making spaces. This makes participation more accessible, flexible, and inclusive. Another innovative approach is the use of AI to synthesize large-scale contributions, organizing diverse perspectives in a transparent and structured way, ensuring that all voices are considered—not only the most influential ones. It is also important to promote hybrid co-creation spaces (both online and in-person), combining technology with human interaction to develop more representative solutions. However, for these formats to be truly effective, it is crucial to ensure equitable access to technology and prevent AI itself from reproducing exclusion. In this way, AI becomes not only the subject of discussion but also a means to enable broad, diverse, and meaningful participation, strengthening more inclusive global governance.

Even laws such as the LGPD, despite their fundamental role in protecting personal data, still face challenges in fully representing and protecting vulnerable groups. This highlights that current regulation is an important step, but still evolving in response to the complexity of society. In this context, there is a significant opportunity for advancement through the integration of new frameworks and international regulations that expand the focus toward algorithmic justice, fairness, and bias reduction in AI systems. Initiatives such as the EU AI Act, the OECD Principles on AI, and UNESCO's guidelines point toward a more mature direction of AI governance by considering social impacts, diversity, and the systemic risks of algorithms. Importantly, the combination of regulation, governance practices, and technical tools can strengthen the development of more fair systems, with better representation of different social contexts in both data and models. Thus, the evolution of AI governance must go beyond data protection, advancing toward a perspective of algorithmic justice and social responsibility in the use of technology.