Loading
Delegates' Area
English
DIGITAL ECONOMY
Impacts of Artificial Intelligence on the working world
While rich countries dominate technology, patents, data and profits related to AI, the Global South is territory for extracting raw materials and supplying underpaid labor without rights. Read an article by Roseli Fígaro, a professor at the University of São Paulo.
In its recommendations on Artificial Intelligence (AI), the Organization for Economic Cooperation and Development (OECD) defined it as: "a set of technologies that seek to make computers do the kind of things that minds can do" (1).
To this end, an industrial park has been set up. The characteristics of this new production chain are very similar to previous patterns: it needs minerals (gold, cassiterite, lithium, cobalt - to name a few - whose reserves are in territories of the Global South, such as Yanomami lands, for example, from where cassiterite is illegally extracted, or cobalt from the Congo, or lithium from Argentina, Bolivia and Chile) and a lot of water to cool the powerful datacenters - the clouds - which use up water and occupy extensive territorial areas.
Data is the input that makes this whole chain move. Data makes algorithms operate and gives machines their own dynamic to reconfigure themselves. Without data, the business model that guides this production chain would not exist. Hence all the clamor for connection, interaction and permanence on social networks.
Materials are also needed for the cables that cross the oceans, connecting the continents or even the satellites that orbit the congested space where the US and China are already fighting over the dark side of the Moon (2). All this is very concrete, heavy, uses natural resources and is polluting.
This infrastructure allows for the existence of the internet, cell phones and all kinds of electronic products, the so-called internet of things and industry 4.0. It is the foundation for the existence of the other operations in the production chain that involve Artificial Intelligence.
Data is another fundamental source. It is inexhaustible because it is produced by social relationships. Data is the input that makes this whole chain move. Data makes algorithms operate and gives machines their own dynamic to reconfigure themselves. Without data, the business model that guides this production chain would not exist. Hence all the clamor for connection, interaction and permanence on social networks. The so-called big data are like mines of data to be processed and organized for specific purposes for all kinds of businesses.
AI feeds on data, without which it cannot operate. We call human data "sensitive materialities": voice, speech, gestures, bodies, eyes, colors, social interactions and activities of all kinds that can be captured, organized and marketed as products - either for algorithmic improvement or to operate in the advertising, journalism, education, medicine and health, commerce, government, personal and national security, even elections markets. We know all this and there are real wars for minerals, water, satellites and data.
However, what drives this whole structure remains invisible: work. They are miners, delivery and app drivers, note-takers who mark and label data, chat mediators, content producers and a range of activities that have not yet been adequately named.
However, what drives this whole structure remains invisible: work. They are miners, delivery and app drivers, note-takers who mark and label data, chat mediators, content producers and a range of activities that have not yet been adequately named. They are part of the great mass of today's working class without rights, they are invisible. Most of them work piecework, a method of payment dating back to the end of the 19th century. There are more than two million workers in Brasil, according to the IBGE (2023).
Brasil is one of the countries with the largest AI data annotation market in the world.
This work can be carried out either by platforms such as Amazon Mechanical Turk or by local outsourced companies - the so-called BPOs (Business Process Outsourcing). There are at least 50 data annotation platforms for AI in Brasil. These workers are the human intelligence that supplies the information and trains the algorithms.
So there's a production chain for infrastructure and there's a production chain for data extraction - we're all part of this data chain, especially those who need to work via application platforms. An app delivery worker, for example, doesn't just deliver food, they deliver data about the city, consumers, commerce, the weather, traffic. These workers are like antennas that collect data for free so that the platform's algorithm can operate.
So we have a new world of work. That factory floor of mobile conveyor belts and fixed working positions, which was transformed by multi-skilling and flexibility, has now been transformed by the explosion of these two orders of organization in the production process. Even in traditional industries, the pace and working conditions are now determined by algorithms that are much less accommodating than the old masters, foremen and supervisors, because they have no empathy. Concentration and control are brutal, and algorithmic work management simulates autonomy and self-will. Hence the false sense of freedom and total deregulation.
Concentration and control are brutal, and algorithmic work management simulates autonomy and self-will. Hence the false sense of freedom and total deregulation.
Global platform companies are not subject to local legislation. The intensification of dependence on Latin American and African countries is evident. They turn a blind eye to constitutional rights and data sovereignty or informational sovereignty. This is the real issue at the frontier of geopolitical disputes. But this banner can only prosper from global efforts to regulate and control these business models.
International bodies, such as the UN, for example, which set up a special commission on governance for AI, have not addressed the issue in the world of work (3); and the ILO's 2023 report did not provide any concrete indications to address the issue (4).
There is deliberate short-sightedness on the central axis of the problem: while patents, data and profits are hyper-concentrated, with a very clear and physical headquarters and address, the countries of the Global South serve as a territory for extractivism, an underpaid, precarious and disenfranchised workforce. An unregulated space.
The persuasive discourse on the new and the technological dazzle prevent us from seeing clearly how these production chains operate and that the only way to regulate them is to regulate labor. The global production chains of contemporary technologies need global governance, but with the participation and determination of the countries of the Global South, which provide a large part of the fundamental inputs for their existence - natural resources, data and living human labor, very much alive, but very poorly treated. But it can be different!
(1) OECD. Recommendation of the Council on Artificial Intelligence. 2019. https://www.oecd.org/digital/artificial-intelligence/ (Tradução própria)
(2) Revista Pesquisa Fapesp, fev. 2019. https://revistapesquisa.fapesp.br/china-pousa-sonda-no-lado-oculto-da-lua/
(3) ONU. Relatório provisório: Governando a IA para a Humanidade. 2024. https://www.un.org/en/ai-advisory-body#:~:text=Co%2Dchaired%20by%20Carme%20Artigas,in%20the%20summer%20of%202024.
(4) ILO. Relatório OIT. 21 de Agosto de 2023 https://www.ilo.org/global/publications/working-papers/WCMS_890761/lang--en/index.htm
Roseli Figaro is a full professor at the University of São Paulo and coordinates the Communication and Work Research Center (CPCT)