At the end of their activities, four G20 engagement groups issued a joint declaration advocating the development and implementation of Artificial Intelligence (AI) in an ethical, sustainable, and inclusive manner. Read an exclusive article for the G20 Brasil website on the importance of the T20 Inclusive Digital Transformation Task Force, written by representatives Bruno Bioni, Jaqueline Pigatto, Louise Karczeski, and Nathan Paschoalini.
Recognizing the benefits and risks that the rapid advance of AI presents in the face of complex global challenges, the groups Civil 20 (C20), Labor 20 (L20), Think 20 (T20), and Women 20 (W20) have teamed up to reinforce the importance of responsible practices related to AI.
The "Declaration of São Luís" presents a shared vision and strategic recommendations for the future of AI within the G20. The document was named after the city where the third meeting of the G20 Digital Economy Working Group took place in June. São Luís was also the site of the "AI Summit: Bridging Boundaries" , a T20 event organized by Data Privacy Brasil , CEWEB.br , CEBRI and Heinrich Böll Foundation , where the basis for the declaration was developed.
This declaration represents an important milestone in the dynamics of the G20 and in creating plural and multi-sectoral policies on such an intersecting issue. With the goal of bringing policy recommendations to the G20 Sherpas Track, the engagement groups sought to align their demands and expectations through a genuinely collective and collaborative process built up over the course of six months.
The consensual declaration reached between groups with different interests strengthens the legitimacy of the document, which was handed over to the Digital Economy Working Group during its final meeting in Maceió, in the hope that the recommendations will influence the final leaders' declaration and the activities at the G20 Summit in November.
Recognizing the need to improve international cooperation on the subject, the São Luís Declaration also refers to AI governance initiatives that have been developed in global forums such as the G20, G7, and the United Nations (UN).
"We emphasize this moment as an opportunity in global governance for international cooperation initiatives aimed at reducing the digital gap between developed and developing countries (...), for aligning our efforts with global initiatives to promote ethical and responsible AI governance, and for encouraging G20 members to advance and strengthen interoperability between AI governance frameworks."
Regarding the UN agenda, the document dialogues especially with the work of the High-Level Advisory Body on Artificial Intelligence and with the Global Digital Pact (GDC) . The GDC and the final report of the AI Advisory Body were published in September 2024 and express the need to improve global monitoring of AI risks and the participation of developing countries in its governance. The documents propose the creation of a Scientific Panel on the subject, a proposal endorsed in the Declaration, which highlights the importance of this panel being established in a Global South country.
Key points of the declaration include decent work, meaningful inclusion, and climate and social justice, taking into account the specificities of different jurisdictions and hyper-vulnerable groups through an intersectional approach. The declaration also endorses combining risk-based and rights-based approaches; and calls for developers and implementers of AI models to publicly report on the capabilities and limitations of advanced AI models and systems.
"(...) improving transparency and accountability, promoting justice and trust through a rights-based approach (...) Assessments on human rights threats and impacts of high-risk AI applications should be reported mandatory, as should providing access to data to qualified researchers and external audits so that such assessments are subject to broader public scrutiny."
The recommendations were based on the data justice approach, which aims to actively fight discrimination and recognize power asymmetries at a global level, guiding the development of AI governance mechanisms that equitably distribute its benefits and risks . In addition, the declaration indicates that the intersectional approach is the theoretical and methodological direction for dealing with the development and implementation of AI systems, guaranteeing the protection of vulnerable groups, especially girls and women, who are increasingly victims of gender-based violence facilitated by technologies.
The groups also recognize the need for a collaborative stance on data governance and representative data sets, emphasizing the need to implement open data policies that promote cultural and linguistic diversity and enable today's workers to adapt to these transformations without compromising the value of their work, by increasing investments in work-integrated educational initiatives such as training programs.
"We recommend incorporating public data generated by citizens to ensure that policy guidelines are informed by and connected to the most local levels of communities, which are often invisible to the state. This can even be done through a 'data spaces' approach within a data economy strategy that is aligned with society's demands and promotes interoperability. We recognize that these policies should encourage equitable data sharing, with appropriate mechanisms for benefit sharing, protecting privacy, and ensuring data security in a legal, transparent, and responsible manner while also avoiding capture by private interests."
The declaration makes an important connection between global governance and local demands, particularly bringing an agenda from the Global South. The four engagement groups responsible for the recommendations-C20, L20, T20, and W20-expect the G20 government representatives to consider them, especially considering the upcoming South African presidency, which presents another opportunity for the Global South to contribute to the AI agenda.
The full text of the joint declaration is available and can be accessed at the link .
Bruno Bioni - Founder and Executive Director of Data Privacy Brasil and co-chair lead of the T20 Inclusive Digital Transformation Task Force
Jaqueline Pigatto - Coordinator at Data Privacy Brasil and deputy co-chair lead of the T20 Inclusive Digital Transformation Task Force
Louise Karczeski - Researcher at Data Privacy Brasil and member of the secretariat of the T20 Inclusive Digital Transformation Task Force
Nathan Paschoalini - Researcher at Data Privacy Brasil and member of the secretariat of the T20 Inclusive Digital Transformation Task Force
