Decolonizing The Machine: From Extractive AI to Abundant Intelligences
Why We Need to Decolonize AI for a Regenerative Future
In the desert outside Phoenix, where aquifers are collapsing and farmers are fighting for water rights, a vast, windowless warehouse hums. Inside, row upon row of servers drink millions of gallons of water every single day to keep from overheating. This is not a steel mill or a coal plant—it is an artificial intelligence data center.
We are rushing toward the edge. Not metaphorically. Not someday. Now.
Capitalism has always been a machine: efficient, extractive, and utterly convinced of its own inevitability. But now, it has given birth to a new kind of offspring—one trained on our memories, fueled by our stories, fed by our attention, and housed in high-security compounds that drink rivers dry and siphon power like gods.
Artificial intelligence, they call it. But make no mistake—it is not neutral. It is the next chapter in a long lineage of empire, and possibly, it is our last.
AI is not saving us. It is not solving climate change. It is not curing cancer. It is accelerating the very system that is unraveling the web of life.
It is concentrating power and capital faster than any other technology in history and it is ushering in an entirely new era of colonization, extraction and control. What we do about it may very well be the defining moment of our species.
Join the conversation on how to resist digital colonization and build abundant intelligences. Subscribe to Regenera for more deep dives on technology, justice, and the future we choose.
Cognitive Colonization
If colonization once came by sea, it now comes by signal—through fiber optic cables, silicon chips and LED screens. It speaks the language of optimization and progress. It learns from our poems and our prayers, our memes and our myths—not to understand them, but to sell them back to us in a flattened, contextless form with a promise attached.
At first they came with steel and scripture. They cut down the forests and built churches instead, saying this is where ‘god’ could be worshipped. Then they drew maps that renamed the land, making everything ‘new’—a mirror of a land across the sea, already colonized, it’s wildlife destroyed. With eyes that refused to see the spirit in the forest or the person in the river they tore down everything that didn’t look like them. They lied. They cheated. They stole. They murdered and tortured and expelled.
And now, they have built algorithms that control our entire world.
The conquest never ended—it simply changed form. From galleons to data centers. From muskets to models. From the theft of gold to the theft of thought and meaning.
Just as colonial schools replaced Indigenous languages with English, French, or Spanish, large language models replace our plurality of tongues with the statistical average of the internet.
What’s being extracted now isn’t just labor or land—but consciousness itself.
A recent Nature study found that frequent use of generative AI tools led to measurable decreases in reasoning and problem-solving skills, especially when tasks required logic and planning (Burtch et al., 2024). The more participants relied on AI, the more passive and error-prone their decisions became.
I can already feel this in my daily life after beginning to use ChatGPT exclusively over Google search, looking for answers for everything from travel recommendations to relational dynamics—the entire field of my life is being consumed by this machine. And because it’s so damn easy and fast, it’s actually harder to expend the energy to think for myself.
Anyone who knows how behavior works can see this is a clear recipe for habit formation and addiction. It’s the very same mechanisms that Instagram and many of the other social media platforms have used to build billions of daily active users who are rotting their brains as they swipe through one outrageous post after another.
Combine this with the neurological fragmentation induced by social media, and we find ourselves in a perfect storm: a population increasingly disconnected from its own mental sovereignty alongside an increasingly polarized political economy and a collapsing biosphere.
As Bayo Akomolafe reminds us: “The times are urgent—let us slow down.”
Energy, Water, and the New Extraction
Today’s large language models are trained on vast oceans of human expression—scraped from the commons, without consent, without reciprocation. This is the datafied equivalent of terra nullius: the colonial doctrine that land not claimed by Europeans was “empty” and thus free for conquest.
But the machine’s hunger is not limited to data. It is physical, material, and it is absolutely ravenous. As I started digging deeper into this, I was horrified. While we have tipped over six of nine planetary boundaries and are struggling with fresh water access around the world, we’re pouring hundreds of billions into energy and water-hungry data centers to further serve the capitalist machine and to make it more efficient.
According to the International Energy Agency, global data centers consumed around 460 terawatt-hours of electricity in 2022—more than the entire United Kingdom (IEA, 2024). That number is expected to more than double by 2030.
In the U.S., data centers already account for 4% of total electricity usage, emitting over 100 million tons of CO₂ each year (Lacoste et al., 2024). Every megawatt poured into the servers is a megawatt not available for decarbonizing the grid.
Water usage is equally staggering. A single AI model, like GPT-3, used over 700,000 liters of freshwater during training (Li et al., 2023). Larger and more complex models would consume many, many times more than this. Cooling data centers requires millions of gallons per day, often sourced from drought-prone regions like Arizona, Texas, and parts of India (EESI, 2023).
“We’re witnessing the digitization of colonial logic,” writes the Abundant Intelligences research team. “A logic that treats land, labor, and knowledge as extractable, ownable, and severable from relationship.” (Lewis et al., 2024)
The AI Now 2025 Landscape Report makes this even more explicit: the AI sector’s environmental costs are systematically externalized onto vulnerable communities, often without their consent, and almost entirely outside binding regulation (AI Now, 2025). The large tech companies are using manipulative tactics to get city and state governments to offer significant tax incentives to build these large data centers, applying enormous amounts of leveraged negotiation on local utility providers for beneficial rates, and passing off the economic and ecological costs onto the local community.
The Oligopoly of the Machine
The AI Now report details how a handful of corporations—Microsoft/OpenAI, Google, Amazon, Meta, Anthropic, Apple, NVIDIA, and major Chinese firms—control the entire AI stack: chips, cloud infrastructure, foundation models, and distribution platforms. Through mergers, exclusive partnerships, and vertical integration, they have created a modern digital oligopoly.
This concentration mirrors colonial monopolies over spice, gold, and trade routes—only now, the commodities are data, compute, applications and attention. The result: smaller actors, public-interest researchers, and local communities have almost no chance to compete or influence AI’s trajectory. Many of the smaller successful AI products either have their features copied by big tech or they get ‘acqui-hired’ and brought into the machine.
Increasingly, governments are becoming entrenched into the pursuit of ‘scale at all costs’ under the narrative of the ‘AI Arms Race’ pinning America against China in pursuit of AGI and the most advanced AI weapons on earth. This toxic narrative is funneling hundreds of billions of dollars into AI that is primarily benefitting a small number of companies and does not have any clear benefit to the general population as a whole.
This capital could be used for creating a highly resilient clean energy grid, for investing in sustainable affordable housing and for supporting regenerative agriculture. Instead, these investments are fueling a demand for energy that has compelled the Trump Administration to sign an executive order ramping up domestic coal mining using growth in demand from AI data centers as justification.
That’s right, we’re going back to coal to feed the machine of AI. And why is that? Because China is doing it…
Hidden Labor and Digital Plantations
What most people don’t understand is that today’s AI is built on the backs of invisible, precarious labor, particularly in the Global South: data labeling, content moderation, annotation, and bias correction. These jobs are often psychologically damaging, underpaid, and intentionally outsourced to avoid labor protections.
This is the digital plantation economy—sustaining the machine’s “intelligence” while hiding the human cost. It is not just the workers serving the machine, but also the humans being surveilled by the machine who are suffering.
AI Surveillance
AI is being increasingly used for corporate and government surveillance, combining computer vision with vast data sets to increase efficiency, identify bad actors and even predict crimes. China’s CityBrain has been using this for years, connecting a digital ID with medical health records, employment data, banking and public transportation—allowing the government to identify individuals through biometrics and to restrict their flow of movement at will.
According to the AI Now report, Amazon tracks warehouse workers down to the minute using AI and will “fire workers if they accumulate more than thirty minutes of ‘time off task’ on three separate days within a year. Time off task includes using the restroom, helping a coworker move a heavy package, or taking a break to cool off or warm up, even when warehouse temperatures are extreme.”
With DOGE compiling the ‘one API to rule them all’ by merging all the major government databases into one platform, we’re likely to experience this level of surveillance and control at a public level across the United States, investigating all of our purchasing behavior, all of our digital communications and all of our movements throughout the country.
The problem is that the underlying LLMs powering these AI systems have been proven to be inherently biased, exacerbating discrimination and prejudice already found in society. As with most parts of the capital machine, “the risks or errors arising from systemically deploying these technologies fall disproportionately on low-income communities, immigrants, and people of color” (Brennan, Kate et. al 2025)
The Epistemicide of AI
What is lost when intelligence is flattened into prediction? When ancestral stories are reduced to tokens? When sacred knowledge is fed to machines that know nothing of sacredness?
Frantz Fanon warned that colonization begins by “depersonalizing” the colonized—stripping away memory, meaning, and sovereignty. Today, we are seeing epistemic erasure on a planetary scale. Indigenous languages, ways of knowing, and worldviews are absorbed by models that cannot hold their context or integrity.
As the Indigenous Protocol and AI position paper explains:
“Knowledge is always relational. It lives in community, in land, in ceremony. AI built within this paradigm must reflect contextual integrity, not disembodied abstraction.” (Lewis et al., 2020)
Instead, the dominant AI paradigm pursues scalability over context. Generality over geography. Efficiency over ethics. In doing so, it extends the empire’s original mission: to standardize, assimilate, and erase.
Toward Abundant Intelligences
The very definition of intelligence in the AI research community models a white supremacist worldview, “a very general mental capability… the ability to reason, plan, solve problems, think abstractly, comprehend complex ideas, learn quickly and learn from experience.” (Lewis et al., 2020)
But what if intelligence could mean something else? What if it were modeled around other cultures, not just colonial ones?
The Abundant Intelligences initiative, led by Indigenous researchers like Jason Edward Lewis and Hēmi Whaanga, offers a radical reimagining. It invites us to build technologies that honor the plurality of intelligences—not just human and machine, but ecological, spiritual, ancestral, and more-than-human.
From the Polynesian Voyaging Society’s wayfinding revival, to Māori-led biodiversity AI projects governed under Indigenous data sovereignty principles, to community-designed language models that serve endangered language revitalization—examples already exist.
“We begin not with the assumption of scarcity, but with the certainty of abundance. Intelligence is not a resource to be extracted, but a relationship to be tended.” (Lewis et al., 2024)
From this perspective, AI could be:
Relational, not extractive
Ceremonial, not transactional
Community-governed, not corporate-owned
Regenerative, not resource-depleting
The AI Now report’s recommendations offer structural pathways to support this vision: enforce antitrust to break up AI monopolies, develop public ownership of key infrastructure, set binding labor and environmental standards, and build global governance models that address cross-border harms.
What Comes Next
To decolonize the machine is to disrupt the myth of inevitability.
It is to refuse the idea that faster is better, that bigger is smarter, that “artificial” means superior. It is to reclaim the right to think for ourselves, to remember for ourselves, to speak in voices that are not flattened for the feed.
For those working at the intersection of climate and technology, this is also a call to:
Fund and support Indigenous- and community-led technology projects
Implement Indigenous data sovereignty principles in climate and AI work
Demand environmental impact assessments for all major AI deployments
Push for binding global labor and environmental standards for AI companies
Because the greatest intelligence may not be in the server rack—but in the seed, the song, the silence.
“Decolonization is not a metaphor.” It is a practice. A prayer. A proposition.
Let us begin—not with answers, but with questions worthy of our ancestors.
Let us build machines that remember they are guests and use them as tools to build a better future, not to serve the slave masters of this world.