However, the growing shift towards an economic system that is dominated by AI investment and technological ambitions will transform the society in ways that lead to new forms of global mass exploitation.
From as early as the Digital Revolution, the accumulation of new technological, especially software, innovations has been associated with common mottos from the companies claiming every new achievement is “making the world a better place to live in”. It has become almost impossible for humans to imagine how the technology that helps them drive to work every day or share pictures on their mobile phones could possibly become an economic or social danger.
The recent revolutionary breakthroughs in AI are no different, except in that they have been subject to early hype from the beginning.
Could AI be a force that elevates and improves the global economy by transforming the job market? Or could it end up elevating and advancing the very aspects of the economic system from which generations have suffered? This piece focuses on two brief accounts of such threats AI poses in today’s world.
The question of control
First and foremost, who owns AI? Though somewhat unacknowledged, it is an established fact that designing and sustaining AI systems, in terms of hardware and software, requires significant resources that span capital, water, electricity, and minerals.
It is no wonder that only huge tech companies such as NVIDIA, Meta, and OpenAI produce global AI-powered products and not the countless universities, institutions, and small businesses that have the technical and scientific capabilities. Hence, how AI is built and controlled is now a matter of company policy.
As with all other forms of trade, intense competition leads to increased monopolisation to either drastically increase AI production or to slow down the increasing call for AI regulation. This is not surprising, as it has been the case with all other forms of trade. For instance, in the same way the most famous multinational energy companies like Chevron leveraged their influence in politics to gain more access to oil imports and avoid tax, the leading AI companies to engage in fierce competition for resources and power.
So far, the increasing demand for silicon, GPU and more advanced electronic chips has become apparent, thus leading to a huge rise in NVIDIA stocks and a clash between China and the United States over the minerals supply chain, which is arguably playing a role behind the scenes in the U.S. minerals agreement with Ukraine and Trump’s interest in Greenland.
Data colonialism
There is a lot at stake when it comes to the software and data requirements of Large Language Models (LLMs) and other large-scale AI systems. It is important to note that the data that is used to train AI systems primarily comes from our activities on platforms such as Instagram, X, Google, and many others — practically all authorised by accepting the terms and conditions of service that we rarely read. Just as in territorial colonialism, where land and natural resources were, and are, exploited, human life, as compressed in patterns of data, is appropriated by a new form of colonialism. This is known as “Data Colonialism“, a term coined by Ulises A. Mejias and Nick Couldry in their book “Data Grab: The New Colonialism of Big Tech and How to Fight Back“.
It is important to point out that the maintenance of such AI systems requires continuous access to more data. While some of the more advanced systems, such as transformer-based models, use synthetic data, most AI platforms require live tracking data. Hence, there is a growing need for continued data collection across all platforms as well as more sophisticated ways of tracking novel features in behaviour that lead to the development of more precise machine learning models.
More importantly, in order for us to continue to use digital services, for personal or professional purposes, we have little choice but to keep agreeing to the increasingly intrusive data collection mechanisms.
A new form of material and intellectual intercourse emerges that encompasses not just goods and services but communication and ultimately new modes of production. Regardless of how relatable this idea is as a theoretical basis of today’s world, it should be acknowledged that a new social, economic, and political order has emerged out of the AI age that is aligned with an increased rate of production and consumption.
The immediate effect is the creation of a complex power relation, in which just as controlling capital bears significant political influence, controlling data amounts to a much more sophisticated form of economic and political autonomy. This data relation is a new and massively asymmetric form of influence in itself, one that focuses all digital activity around maximising data extraction and monitoring in order to maximise profit.
The empires of AI
An additional, and more recent, aspect of data colonialism is the competition over achieving Artificial General Intelligence (AGI). AGI is meant to surpass human intelligence across all cognitive tasks, including natural language, vision, and even logic and reasoning. The ultimate ambition of companies such as Google’s DeepMind, Meta, and OpenAI is to build AGI that learns from scratch like human beings, but achieves an ideally optimised performance in virtually all areas of activity through integrating its skills and training. Unlike most of the current AI agents that stop learning after their training period, AGI is meant to keep learning from experience and live data, meaning increased access requirements to tracking data sources and empirical engagement with the environment.
The growing data requirements of AGI are somewhat predictable, but the hardware and energy requirements, as well as their potential environmental overhead, are far from obvious. Yet the few companies and institutions such as OpenAI and Google’s DeepMind have already invested heavily in projects that ultimately aim at developing AGI. We are observing an increased attempt by corporations to secure funding, data, and less regulation at a global stage in an already chaotic and economically bankrupt era.
The competition for AGI and its colonial nature, particularly by OpenAI, has been recently explored by Karen Hao in her “Empire of AI: Inside the reckless race for total domination”, likening tech companies to empires due to their expanding magnitude and accumulated economic and political power. She has attempted to raise awareness of the fact that this power is not just a consequence of capitalism, but also one fuelled by the pursuit of AGI, and that it could be the ultimate scientific achievement of mankind. Much like colonialism, which operated not only through economic and military means but through an ideological obsession that, however wrong, could logically justify and elevate a cause at any cost.
Just like any innovation in world history, AI can be a force for good as much as it can cause havoc with devastating consequences in our communities. For now, it seems that the race is fierce, and that AI research and development cannot be stopped or redirected from the path it has taken. However, just like the most successful struggles against colonialism, raising awareness and criticising the existing social relations should be the first step in developing an equitable society, in which AI can benefit the masses and the individual.
