Technology as Something We Can Build for Ourselves

For me, technology has always been a thing that I can create with and create within.

I’ve always had a kind of fascination with setting up technical things, even though I don’t know programming, and even though I’m not a coder. I found coding interesting in school, but I never quite got a grasp on it. AI is kind of changing the necessity of that, to be honest, but that’s a whole side conversation.

The point is: I’ve always seen technology as something I can be an active participant in.

And I’m realizing, especially in the discussion around AI, that a lot of people talk about AI as if it’s something happening to them. Not something they can have any participation in. Not something they can co-create, adapt, modify, or make for themselves.

There’s a larger thread of that with technology overall. People complain about big tech and social media as if they are being forced to use these platforms. As if they have no power to create their own platforms or build their own digital spaces.

And I get it. It’s not simple. But I’ve always seen technology as a sea of alternatives. What else can we do? What can I do to solve the problems I see with technology?

That’s also how I viewed AI.

When AI first came on the scene, I tried a couple of models. I made up a couple of things just to see what it could do, just to see what the capabilities were, mostly out of interest and curiosity.

And I found that it didn’t do some things well.

I didn’t like using it for creativity, or as a primary creative tool. I didn’t like writing stories using AI. I still like the process of creating slowly or manually. That’s one of the things I learned early on: for me, at least, AI would never replace my creativity.

I can add it to my workflow. I can use it to embellish things, make small changes, create assets, or help with certain tasks. But overall, the fundamental creativeness still comes from me, and I prefer it that way.

But in the larger conversation, the framing has often been: “They’re forcing AI on us.” Or, “AI is unethical.” Or, “It’s trained on all our data.” Or, “It’s using all this water.”

And granted, these things are true. Very true.

AI has a massive ecological, human, and social cost. It is damaging to many people’s livelihoods in very real ways.

But some people use that to say, “AI is fundamentally bad.” Or, “All AI is terrible for the environment.” Or, “AI should not exist”

And from my experience of using AI, I don’t think it’s that clear-cut. It’s not simply “AI is bad” or “AI is good.” It’s not “AI is evil” or “AI is salvation.” Like every technology, artificial intelligence has nuance.

Early on, in the same way that I tried to adopt open-source, decentralized social media as a reaction to centralized, closed-source social media, I started thinking about AI in terms of agency.

With social media, I eventually failed at that switch because people didn’t want to move over. People generally don’t want to switch from these platforms. And if I need social media, then I have to be where the people are. Otherwise, it’s not really social, and that’s kind of the point of it.

But with AI, I had a similar question:

How can I take some ownership or agency over the AI I’m using? How can I create a more ethical way of using AI?

And for me, it’s not as simple as searching for “better” AI providers.

Some people use Claude over OpenAI because they say Claude is more ethical. But for a lot of people, cloud AI from these large corporations is the only idea of what AI is.

They don’t know about open models. They don’t know about open-source technology and software. They don’t know that a lot of the fundamental technology underlying these AI systems comes from open research and open-source software.

What that means, broadly, is that much of the code, research, and tooling is free and available to the public. People can study it, use it, and build with it.

The underlying technology was created and given to the world by researchers and open-source communities. Then big corporations took that technology, created large models, closed them, made them private, charged people for access, and ran them in massive data centers with massive ecological costs.

But the fundamental technology is not inherently corporate.

As a matter of fact, right now, many people have devices that can run some of these models locally. There are smaller and smaller, more capable open-weight and open-source models coming out that people can run on their own devices without needing to connect to the internet.

That means no constant cloud dependency. Less or no corporate data harvesting. No reliance on water-intensive data-center cooling for every query. And it is possible to run these systems on solar or other renewable energy, removing a lot of the traps and concerns of big tech and big AI.

For me, open models solve, or at least directly challenge, one of the major ethical problems of AI.

These AI models are trained on internet data: people’s artwork, pirated books, massive data troves of media and text from across the internet. And the internet is something the whole world created.

So I think it is fundamentally unethical for companies to use the world’s publically available data, then close-source their models and gatekeep access so that only a select few who pay them can use those models.

These models would not exist without our collective data.

I believe all models trained on public and collective data should be open, accessible, and treated as part of the commons, or they should not exist. That’s my ethical position.

The legality of these things is still very much up in the air, and some people are still battling it in court. OpenAI and others have been allowed to get away with it under the law because the act of making these models can easily be argued as fair use or transformative work.

I don't disagree that the act of making LLMs using transformers is transformative. But my stance is my stance:

If these models are going to be created using our data, then they should be open. Everybody should have access to them. Everybody should be able to run them, train them, modify them, and use them for whatever they want.

So I started exploring how to run open models on my own hardware.

Eventually, I got to a point where I could create a ChatGPT equivalent. It probably isn’t as smart. Actually, it definitely isn’t as smart. But it is able to do a lot of what ChatGPT does.

And it’s running on my own PC, on my own hardware.

My data stays mine. My data remains private. It doesn’t rely on water-intensive data-center cooling. It doesn’t have any the massive social and ecological cost during its operation. And when I’m able to connect it to solar, I can make it much lower-carbon to operate and use.

For me, that solves a lot of the problems of AI. Or at least, it directly addresses many of the qualms and concerns I have with AI.

So I set it up.

I’ve been exploring it, updating it, upgrading to newer models, and testing these models for different purposes.

And of course, it’s less convenient. Of course, it’s not as performant as the big cloud-based systems. But that is a constraint of the hardware, not necessarily a constraint of the models themselves. If I had a better GPU, I could have better performance.

And I think that is fundamentally the point.

The point is that we can create our own systems.

I built this PC two years ago with a bonus from my last job, my only nine-to-five. I built it myself, and it was fun, interesting, and rewarding. It was my first real PC build. After building a few computers in school, and after watching videos online, I just put the thing together.

And there was something deeply rewarding about that.

Setting up these systems can be difficult. It can be frustrating. Sometimes you follow all the instructions, put in the terminal commands, do everything you’re supposed to do, and it still doesn’t work.

That part is annoying. Deeply annoying.

But eventually, you find the answer. Eventually, it works. And there’s a sense of fulfillment in that. A kind of personal investment. A kind of meaningfulness.

A lot of modern technology doesn’t have that anymore.

The internet, social media, and many modern AI services don’t give people that sense of making and building something themselves. Technology is increasingly made by big corporations in the United States and other so-called “first world” countries, and then sold to us.

It uses us. It farms our data from the Global South, from so-called “third world” countries — and I really hate those terms — and then sells that data back to us as expensive new technology.

These companies exploit us digitally. They exploit our collective data, and then sell it back to us as products we have to pay more and more money to access.

But there is an opportunity here.

There is an opportunity for us to take back sovereignty over our data and our technology.

We can build our own systems. We can create our own networks. We can create our own technological capacity.

Open-source models and software are created collectively. Strangers on the internet contribute code, build tools, and work on projects in their free time. In that same way, we can come together as caring people, as Jamaicans, as marginalized people, as exploited and disenfranchised people, and make our own systems.

Especially from a social justice perspective, we know these AI companies do not care about ecology. They do not care about the preservation of Indigenous, Black, and brown cultures. We know Caribbean data is underrepresented in a lot of these AI models.

All of these new technologies are being made without us at the center.

Facial recognition technologies often do not reflect how we look. Language models often do not reflect how we speak, how we think, where we are from, or who we are.

Because they are made largely from a white, Anglo-centric perspective. And the internet itself is heavily Anglo-centric, which is where a lot of the data comes from.

So there is an opportunity for us to make our own thing.

To invest in our own technologies.

To create our own capacity.

And yes, a lot of these open-source technologies are still made by corporations. But they are open source, which means we can change them. We can build on them, adapt them, and use the technologies available to us to create something that serves our own needs.

We don’t even need to know how to code to start.

I don’t know how to code, and I was able to set this up.

Imagine what we could create if we came together and actually made our own thing.

Ultimately, I think we need to stop thinking about technology as this evil thing that is simply happening to us.

We need to stop treating AI as only something that is taking away our jobs, destroying our livelihoods, and damaging the environment, while we stand around as passive bystanders to our own exploitation, our own sidelining, and the destruction of our environment.

Those harms are real. But passivity is not the answer.

We need to take claim of technology and create something that is actually beneficial for us and for ecology.

After talking to a few people who showed genuine interest, I realized that this hobby thing I made for myself could actually be useful to others. I was surprised by that.

So I’ve decided to make it available for people to use and test. I want to get feedback on how well it performs, see how it handles multiple people making queries, and explore offering a service that provides open intelligence to people in my circle and my community.

That way, they don’t have to contribute to AI systems that exploit them, destroy environments, drain water from communities, pollute rivers, and concentrate power in the hands of corporations.

They can use AI that is more ethically aligned.

AI that they can actually train themselves.

A lot of these models can be guided with information. You can give them more Caribbean context. More queer context. More information about who you are and your part of the world.

And you can do that without giving that data up to corporations that would gladly use it to exploit you further, make money from you, and train their own closed models.

A lot of these problems are addressed by hosting these systems ourselves.

And it is not just AI.

I also have my own cloud system that I set up before this. I figured out how to set up Nextcloud, which is an open-source alternative to Google Workspace. I currently offer it to my clients, but I’m starting to think I could offer it to people in social justice movements adjacent to me too.

They could have their own cloud storage and collaboration software that is not tied to companies they are not aligned with. Not tied to big corporations they oppose. Something cheaper, or free, or locally supported.

A few people I’ve spoken to have said they would rather pay me for access to something like this than pay a faceless corporation.

And honestly, that makes sense.

There is a kind of local queer economy that could be developed around hosting our own tools and providing for each other using the technology we already have.

Using the master’s tools, yes — but using them to create new ways for technology to exist.

Better ways.

Ways for technology to exist in harmony with our environment and our people. Ways that make our lives easier and our work more impactful, without carrying all of the same negative consequences.

It is wild how many people are already using AI in the NGO space, in charity work, and in social justice work, while having no real idea that alternatives exist.

We hear about all these issues with AI and feel conflicted. We feel like there is nothing we can do, in the same way we feel conflicted about social media.

But there is something we can do.

We can stop treating technology as something only corporations get to make.

We can build our own.