AI and the Cotton Gin

Kaleb Evans-Lao

When I think about the new AI data centre planned for Makarewa, I think about how I would have responded ten years ago to the idea that a revolutionary new technology was being built in our own backyard—one that promised to completely change our way of life for the better. I would have been excited. I’ve always had this notion that new technology automatically meant progress, change, and better living standards.

But in recent years, that optimism has waned. And I would argue it isn’t just my own attitude towards technology that has changed, but the broader cultural consensus as well. We seem far removed from the technological optimism of The Jetsons, Star Trek, or even that meme image frequently accompanied by captions about “the world with/without _____.” Instead, we seem more in line with the technological portrayals of I Have No Mouth, and I Must Scream and Terminator—or, more recently, Oppenheimer.

We’ll come back to techno-optimism shortly. First, I want to make a comparison between today’s rise of generative AI and the invention of the cotton gin.

Both demonstrate that technological progress does not happen in a vacuum. A new device or method enters a world shaped by existing priorities and incentives, and people put it to use accordingly. In both cases, that has meant furthering wealth at the expense of others.

Ultimately, we should be more critical of new technology. When something promises to boost productivity, we should ask: who does that increased productivity primarily benefit?

First, the history of the cotton gin, short for “cotton engine.”

Invented by Eli Whitney in America in 1794—though Whitney’s involvement in its invention is disputed, it is certain he was instrumental—the cotton gin revolutionised cotton production. Before the gin, processing cotton involved picking sticky seeds out of cotton fibres by hand. It was slow, arduous work performed by enslaved people on American plantations. One enslaved worker could de-seed roughly one pound, or half a kilogram, of cotton per day.

With the gin, that increased to 50 pounds, or 22.6 kilograms, per day for two enslaved workers operating the machine. It was undoubtedly a substantial increase in productivity. The reaction from northern abolitionists, enslaved people, and Whitney himself was that the machine might help bring an end to slavery: theoretically, plantation owners would no longer need as many enslaved people to pick seeds from cotton.

Unfortunately, the opposite occurred.

Before the gin, cotton and other plantations reliant on slave labour were seen by some as a sunset industry, with the cost of keeping enslaved people gradually outscaling the revenue generated by the crops they produced. The idea that slavery might therefore be phased out economically has been disputed by Reconstruction historian Paul Finkleman, who wrote in 2013: “Slaves were a profitable investment before the cotton gin and an even more profitable investment after its invention.”

Nonetheless, the belief that slavery might eventually become economically unviable was not an absurd one at the time.

What is indisputable, however, is that the cotton gin made the cotton industry—and therefore slavery—significantly more profitable. More enslaved people were moved from other plantations to those producing cotton, irreparably separating families and communities. As cotton became increasingly profitable, mass cultivation also degraded soil quality on eastern plantations, and cotton plantations—along with the institution of slavery—expanded further west.

The cotton gin helped drive that expansion.

Now to generative AI.

I don’t think I need to tell a Salient reader why AI is abhorrent. But I do want to point out an aspect of its harms that you may not have immediately considered.

Large language models and generative AI have created job opportunities. But many of these jobs exploit vulnerable people who are desperate for work, offering little and inconsistent pay. Worse still is what some workers are required to do for that pay.

AI companies often offer remote work to people desperate for any form of employment, or who cannot enter traditional workforces. The work can involve menial tasks such as sorting and labelling data to train and improve AI systems for companies like Meta, OpenAI, Microsoft, and Google. In particular, workers tasked with filtering content before it enters AI databases can be required to regularly view horrific material that I will not divulge here.

These “human-in-the-loop” jobs have existed—and exploited people—since social media conglomerates first needed humans to filter their content. AI has only furthered this unethical practice. No matter how sophisticated generative models appear to become, there remains a constant demand for menial human input to keep them functional.

So, what do these two examples tell us about technology?

In both cases, increased productivity came with a promise of broader social benefit. In reality, it further entrenched exploitative practices that were already in place.

And who benefited?

In the case of the cotton gin, it was American cotton barons and plantation owners. The United States eventually supplied three-quarters of the world’s cotton, while Britain’s industrial textile mills created enormous demand. Plantation owners grew incredibly rich.

With AI, the parallels are easy to see. Figureheads such as Sam Altman have benefited from the massive allocation of funds towards the AI industry. Ultimately, he and the companies that rely on exploited workers to keep their supposedly “autonomous” machines functioning reap the rewards of other people’s gruelling labour.

Now, techno-optimism.

It’s a loosely defined term, although the idea has arguably existed for many years. Generally speaking, techno-optimism is the belief that technological “progress” is the best—or sometimes the only—way to solve existing problems.

The term is sometimes applied to approaches to climate change in which environmental problems are imagined as solvable through emerging or even fictional technologies: if technology progresses far enough, the thinking goes, it will solve the climate crisis.

Looking at you, solarpunk.

Techno-optimism is a fundamentally capitalist idea. One of the longstanding defences of capitalism is that it breeds innovation; new methods and machines have been touted as one of capitalism’s great strengths since the Industrial Revolution.

The validity of that claim aside, presenting technological innovation as inherently good relies on the assumption that innovation itself produces positive outcomes. I hope my spiel about the cotton gin has shown that technological gains in productivity can coexist with—and even enable—devastating social consequences.

And techno-optimism is everywhere. If you’ve ever watched a glossy product-presentation video, you’ve seen it in action.

Think NEOM and its idiotic “The Line” project in Saudi Arabia. Facebook’s “Metaverse.” Whatever Elon Musk has contributed to the transportation industry. And, of course, AI.

Everywhere, tech giants push the idea that technology is the way forward, all the way towards transhumanism and the technological singularity—which is a whole other can of cyber-worms.

My overall point is that we need to think about technological progress more critically. Not in the “I won’t vaccinate my children” sense—that isn’t critical, it’s ideological. Rather, when we encounter a technological breakthrough, we need to examine how it will actually operate within the society we already have.

In the case of generative AI, it may be capable of doing good. But when it is placed within a system that prioritises profit over wellbeing, it is inevitably going to exacerbate problems within that system while creating new ones of its own.

So, finally, if someone tells you to “just embrace the technology,” or offers some other phrase along those lines, tell them the history of the cotton gin.

Then ask them: who did that new technology actually benefit?

Further reading (and watching):

“Seeds of a Lie,” a Youtube video by Extra History providing a well-produced breakdown of the history of the cotton gin.

“The Artificial Intelligence Illusion: How Invisible Workers Fuel the ‘Automated’ Economy,” an ILO article highlighting how AI is not fully “autonomous”.

“Kenyan Workers With AI Jobs Thought They Had Tickets to the Future Until the Grim Reality Set In,” a CBS article. There are many pieces on this topic, but this provides an extensive look at the harms associated with AI work.