Disclosure: The views and opinions expressed here are those of the author and do not reflect the views and opinions of crypto.news’ editorial team.
Currently, AI is everywhere, impacting everything from productivity enhancement in workplaces to managing complex personal issues. While this innovation brings advantages, it falls short in making a significant difference in the world’s most disadvantaged areas.
Summary
- Centralized AI fails the Global South, perpetuating bias, undermining data sovereignty, and creating opaque, unaccountable systems that contradict the UN’s Sustainable Development Goals.
- Decentralized AI — utilizing federated learning and blockchain — presents an inclusive, secure, and transparent alternative that fosters local data governance, responsible management, and tangible applications in climate action, healthcare, finance, and conservation.
- The way forward requires transitioning from corporate AI to an open, decentralized framework, ensuring AI ethically contributes to global development by embedding principles of inclusion, sovereignty, and accountability into its design.
The United Nations Development Programme is unwavering in its quest for the 17 Sustainable Development Goals aimed at eliminating poverty, advancing climate action, and fostering equitable growth by 2030. Given its existing global applications, AI could be seen as pivotal in enhancing inclusivity and global development. Nevertheless, the current centralized AI framework is riddled with challenges such as privacy issues, exorbitant costs, and limited accessibility, greatly diminishing AI’s potential for positive influence.
This centralized nature of AI only serves to exacerbate existing power disparities, hindering AI’s ability to effect positive change in the Global South due to inherent biases, loss of data sovereignty, and a lack of transparency. To harness AI effectively as a mechanism for global development, a shift from the corporate, centralized model to one emphasizing inclusion, sovereignty, and accountability is essential. Decentralized AI is the solution.
The centralization paradox
Though AI has been employed to tackle issues ranging from climate change to healthcare, its development remains largely controlled by a few major tech corporations, whose systems are not ideally suited to meet the unique challenges posed by the 17 SDGs established by the United Nations. The issue isn’t technological; it’s about governance. The traditional model of AI development presents three significant barriers to achieving genuine developmental impacts.
Centralized models predominantly rely on data sourced from a limited number of developed regions, neglecting those in the Global South. Research indicates that these models perform poorly in diverse settings, such as diagnosing illnesses or assessing financial risks. Without proper training, they can lead to systematic misidentification, denial of essential services, and the exacerbation of socioeconomic inequalities, jeopardizing SDG 10, which aims to foster social, economic, and political inclusion for all.
These systems also necessitate the accumulation of highly sensitive local data, from patient records to financial or criminal history, on distant corporate servers, which are vulnerable to breaches because of their centralized nature. This data extraction strips governments and organizations of their data sovereignty rights, undermining SDG 16, which champions peace, justice, and robust institutions, and threatening the security of information stored away from local servers. This practice has also prompted the emergence of sovereign AI technologies in countries like Singapore and Malaysia, as they vie to maintain data sovereignty.
The critical question remains: when an opaque, poorly understood AI system makes a grave error or misjudgment regarding policies that could impact millions, who is accountable? The “black box” characteristic of centralized AI systems complicates auditing processes, such as determining aid distribution or risk assessment, and makes it exceedingly difficult and ethically troubling to establish responsibility in high-stakes development scenarios. This opacity threatens to hinder all 17 SDGs.
The only path forward to align AI’s capabilities with the ethical demands of international development is a substantial move away from corporate, centralized AI training toward methodologies grounded in inclusion, sovereignty, and accountability.
Decentralized AI: A two-pronged solution
Decentralized AI, bolstered by federated learning and blockchain technology, is emerging as a viable solution. The SDG Blockchain Accelerator Programme, strategically led by the UNDP and supported by partners such as Blockchain for Good Alliance, Stellar, FLock.io, and EMURGO Labs, validates this concept by spearheading decentralized AI efforts that empower communities in the Global South.
Federated learning operates by training shared models across various decentralized devices, preserving local data privacy. Initiatives in the Latin America and Caribbean region utilize this technology to collaboratively develop predictive AI that accurately forecasts climate-related risks while safeguarding local financial and demographic data on local servers. This framework enables fair and efficient payouts to climate-vulnerable farmers and women-led businesses, fulfilling both SDG 13 (climate action) and SDG 5 (gender equality).
These federated learning operations are enhanced by blockchain technology, which replaces a single corporate intermediary with a transparent, immutable governance framework. This establishes a collaborative infrastructure and reinstates essential accountability mechanisms. In Liberia, smart contracts and decentralized AI are being implemented to ensure transparent distribution of payments and aid, and in Kenya, decentralized AI resolves payment inconsistencies for local businesses, fostering economic growth and trust in public institutions, as emphasized in SDG 8 (decent work & economic growth) and SDG 10 (reduced inequalities).
Additional applications of decentralized technology supporting the SDGs include a blockchain-based NFT developed by Cambridge University and UNDP Rwanda for mountain gorilla conservation, and secure hospital records in Africa that grant patients the autonomy to control who accesses their medical information, addressing SDG 15 (life on land) and SDG 3 (good health & wellbeing).
A call for architectural responsibility
AI as a technology holds tremendous potential, but its primary challenge is governance. The centralized, proprietary model fundamentally contradicts the principles of inclusion, sovereignty, and accountability represented by the UNDP SDGs. Current initiatives show that a viable, ethical, and scalable alternative is possible.
The global development community must now prioritize funding the transition to open, decentralized AI frameworks over corporate technologies that hinder progress. We must shift from being passive consumers to active custodians of intelligence that foster a sustainable future for the most marginalized communities worldwide.
