-
CENTRES
Progammes & Centres
Location
AI’s global boom is powered by the Global South’s labour and land—but its priorities remain sidelined in a race defined by corporate greed and power.
.jpg)
Image Source: Getty
There is no doubt that the rapid development of Artificial Intelligence (AI) has been a huge boon for humanity in recent years. Finding applications in virtually every facet of life, AI has accelerated progress in a variety of domains on an unprecedented scale. However, this groundbreaking progress has come at a great cost - not just in terms of how a model includes or excludes certain populations, but also in the costs extracted by expanding infrastructure. The proliferation of data centres across the globe has sparked widespread protests and discontent amongst local communities. Furthermore, labour-intensive tasks such as data labelling—critical for the development of AI models—have had a detrimental psychological impact on individuals involved with the process. This highlights the need to rethink the development paradigm in the age of AI. As intense competition persists among global hegemons, it becomes pertinent to ensure that the global prerogative is defined equally through the effective participation of the Global South and its development priorities.
The exponentially increasing demands for AI data centres by big tech corporations and the concomitant land, energy, and water requirements have prompted them to find data centre hubs in the Global South countries. In 2023, Google’s plan to develop a data centre in Uruguay faced massive backlash from the local population, especially considering the country was suffering its worst drought in over 70 years, and only a small section of the population had access to clean drinking water, a large portion of which risked getting diverted to the data centre.
Labour-intensive tasks such as data labelling—critical for the development of AI models—have had a detrimental psychological impact on individuals involved with the process.
Chile has 22 data centres, with 16 of them in Santiago. This has led to severe stress on the country’s water resources and environment. Although Google’s plan to build a second data centre was approved by the Chilean government in 2019, it was later suspended in 2024 by Santiago’s environmental tribunal owing to a series of demonstrations by the Socio-Environmental Community Movement for Land and Water (Mosacat), a local activist organisation. Other corporations, such as Amazon and Microsoft, have also faced significant protests in the country. A similar situation exists in Brazil, where five of the 22 planned data centres—including one by Chinese tech giant ByteDance—are located in cities with severe and protracted water scarcity issues.
As intense competition persists among global hegemons, it becomes pertinent to ensure that the global prerogative is defined equally through the effective participation of the Global South and its development priorities.
Moreover, the damage caused by the widespread proliferation of data centres is not merely limited to the Global South, but has now spilt over to global hegemons themselves. According to a report by Data Center Watch, about US$64 billion worth of data centre projects in the United States (US) have been blocked or delayed due to domestic opposition. Existing data centres, such as xAI’s Memphis facility, are also facing protests due to accusations of enhanced air pollution in areas like Boxtown. A plan to build Europe’s largest AI data centre in London’s green belt is facing similar backlash from the local community.
AI models rely on data labelling to enhance accuracy and filter out explicit content. To maximise profits, big tech corporations outsource these tasks to Global South nations with cheap labour, often employed at exploitative rates. For instance, to remove toxic content such as violence, sexual abuse, and bias from ChatGPT, OpenAI employed Kenyan workers earning less than US$2 an hour through its outsourcing partner, Sama, which operates in Kenya, Uganda, and India. In addition to being vastly underpaid, the traumatic nature of the work has had a severe psychological impact on multiple Kenyan workers, which eventually led to Sama cancelling its work for OpenAI eight months earlier than it was planned. Furthermore, in 2024, almost a hundred Kenyan AI workers sent an open letter to former US President Joe Biden stating, “Our working day conditions amount to modern day slavery.”
The exponentially increasing demands for AI data centres by big tech corporations and the concomitant land, energy, and water requirements have prompted them to find data centre hubs in the Global South countries.
Similarly, the ongoing crisis in Venezuela over the past few years has made it a hotspot for the outsourcing of data labelling through companies such as Appen. The economic desperation of Venezuelan workers has been actively exploited by corporations that have systematically lowered their wages, in addition to suspending accounts and discontinuing programmes.
The situation persists across the Global South with exploitative practices, including reduced wages, subjecting employees to hazardous working environments, and a general lack of transparency. Furthermore, this also involves labelling providers pushed to resort to child labour to fulfil their ever-increasing demands. India, being home to the largest number of online freelance workers in the world, is a particularly easy target for these companies.
Regardless of the complexity of the problem, multiple solutions are always possible. This includes innovations prioritising computational efficiency, like DeepSeek, exemplifying the fact that focusing on efficiency and open weights can serve as a viable avenue for the future of AI models. Moreover, alternatives such as Small Language Models have also shown promise, paving avenues for neutralising several of the above concerns, particularly since the possibility of offline functionality in some cases can render them free from dependence on cloud computing infrastructure. However, these solutions have not received enough attention, with the majority of capital expenditure by big tech corporations such as OpenAI, Microsoft, and Google instead focused on building massive infrastructure solutions required for global AI expansion.
Solutions have not received enough attention, with the majority of capital expenditure by big tech corporations such as OpenAI, Microsoft, and Google instead focused on building massive infrastructure solutions required for global AI expansion.
To participate effectively in AI-led global development, the Global South must focus on developing indigenous models and more efficient alternatives better aligned with its interests, while strengthening labour laws and enhancing transparency across AI supply chains. While advancing AI development has its rewards, it must not compromise human dignity and progress. The upcoming AI Impact Summit to be hosted in New Delhi in 2026 serves as an opportune moment for India to assume leadership on this critical issue and prevent the Global South from falling victim to skewed power balances in the AI race. It can achieve this by taking a firm stance on the issue, fostering ideological and economic cohesion among like-minded nations, and furthering the interests of the Global South, which continues to suffer within the exploitative grips of corporate behemoths.
Prateek Tripathi is a Junior Fellow at the Centre for Security, Strategy and Technology (CSST), Observer Research Foundation
The views expressed above belong to the author(s). ORF research and analyses now available on Telegram! Click here to access our curated content — blogs, longforms and interviews.
Prateek Tripathi is a Junior Fellow at the Centre for Security, Strategy and Technology. His work focuses on emerging technologies and deep tech including quantum technology ...
Read More +
PREV
