The author of ‘Empire of AI’, a book about OpenAI and its CEO Sam Altman, discusses technology, power and sovereignty.
Alessio Giussani is the editor-in-chief of the Green European Journal . He was formerly contributing editor of Eurozine (Vienna) and a freelance journalist based in Athens, Greece.
Karen Hao is an award-winning journalist covering the impacts of artificial intelligence on society. She co-hosts the BBC podcast The Interface and contributes to publications including More Perfect Union and The Atlantic . She
Cross-posted from Gre en European Journal
Alessio Giussani: Several metaphors have been evoked to describe the current model of AI development – a parrot or a mirror , for example. Why did you choose that of empire?
Karen Hao : I think those metaphors are all extremely useful. I focus on empire because I’m not just looking at the technical artefact itself and what happens when it’s deployed in society, but also at the production of this technology and the actors who consolidate an extraordinary amount of economic and political power to produce it.
There are four main reasons why I call this system an empire. First, these companies lay claim to resources that are not theirs: individuals’ data, as well as the work of writers and creators. Second, they exploit an extraordinary amount of labour, and that goes for the workers they exploit within the production process and the supply chain of these technologies, as well as the workers whose labour rights are eroded when the technology is deployed.
The third aspect is information control. A major facet of empires’ expansion is the ability to filter all knowledge production and information through the lens of their imperial agenda, and this is precisely what AI companies are doing. They control and censor the science that gets produced by bankrolling most AI researchers, and they are building an information technology that they seek to make the central portal through which everyone else engages with the world, including in education.
Lastly, they wrap all this in a civilising mission. They say that they’re engaging in this work because ultimately they are trying to bring humanity a technology that will elevate us to a utopian state, and if the bad guys – which usually means China – develop it before us, then we will instead descend into a dystopian state. This is a kind of religious narrative that was also common among empires of old.
The AI empire, as you are describing it, is not limited to a company; it’s a system, a logic. What lies behind your focus on OpenAI and on the figure of its CEO, Sam Altman?
One answer is that people need stories to help make concrete what would otherwise be an abstract idea. But OpenAI is also particularly interesting because it made all of the major decisions that changed the norms within the industry, pushing it towards empire-building. It was OpenAI that decided to scale at all costs, to work with large language models, and produce a chatbot-like interface for them. Pretty much everything that the public understands to be associated with AI today was first decided within OpenAI’s walls.
By telling that history, it becomes much more obvious that AI is not inevitable, because you start to realise there were real people making those decisions based on their values, their worldviews, their personal vendettas – whatever it is that actually shaped the trajectory of AI development. Sam Altman’s ideologies are imbued in the model, not just in an abstract sense. You often hear the truism that models reflect their creators, but I wanted to show how this actually happens. It’s about the companies, the people, and the economic and ideological environment they are part of.
As these technologies grow more powerful, asserting control over their deployment is becoming an existential question. We saw Big Tech bowing to Trump at his second inauguration, but we are also seeing AI CEOs increasingly at odds with the government. Who will end up having the final say – the state or private companies?
There is an ongoing battle for who’s going to call the shots. I see the relationship between Washington and Silicon Valley as an alliance of convenience, but a very fraught one. They represent two different models of empire, each with its own imperial agenda. Both are trying to expand and quite literally take over more land from other countries, but the Trump administration sees AI companies as tools in the hands of the state’s imperial ambitions. Meanwhile, Silicon Valley sees the Trump administration as a tool of their imperial ambitions, but they’re not actually on the same page, because each of them wants to be the supreme entity that comes out on top.
There have been moments when the US government has demonstrated that it’s still on top – when Trump and Musk collided, Musk got booted out. However, when the Department of War clashed with Anthropic, the company showed that it was the stronger power in that instance: Secretary of Defence Pete Hegseth evoked the nuclear option, threatening to declare the company a supply chain risk as the ultimate stick that he could whack to get Anthropic in line, and the company still didn’t get in line. Quite embarrassingly, the Department of War has become deeply dependent on a private actor that is unwilling to cooperate on the government’s terms, and the government can do nothing about it.
There needs to be accountability for both of these imperial actors, and that accountability will not come from either of them; it has to come from the people. From bottom-up, grassroots organising, which we are beginning to see blossoming all around the world, including in the US. As impervious as the alliance of convenience may appear, it is actually extremely fragile, and there are opportunities to accelerate its demise by applying more pressure at its pain points.
The state of AI development looks like a competitive race to the bottom between rival companies and global superpowers. What role has geopolitical competition with China played in bringing AI to its current predicament?
China is a card that Silicon Valley has played for a decade, pretty much ever since it got kicked out of China. But when you look at the track record of this argument, you see that what has happened is consistently the opposite of what Silicon Valley said would happen.
They said that by not regulating Silicon Valley companies – and by regulating China through export controls – American technology would dominate worldwide. Yet TikTok is now the most popular social media platform, and Chinese open-source AI models are becoming more and more popular around the world, including in the US. They also promised that Silicon Valley dominance would have a liberalising effect on the world, but that hasn’t come to pass either; their effect has been deeply illiberalising. There has been an extraordinary erosion of democratic institutions and a rapid acceleration of democratic backsliding, aided by Silicon Valley platforms and AI technologies.
Instead of looking at AI as a global race between the good and the bad guys, we should be thinking about what vision we have for society, what kind of AI technologies would support that vision, and what resources we need to bring it about. We should be competing not in a race Silicon Valley defined, but on the ideas and values we want to uphold. And I believe Europe has an opportunity here to show the world a third way, different from those of the US and China, that resets the rules of the game.
The European Commission has recently adopted a tech sovereignty package as part of Europe’s effort to chart its own technological path. How should we articulate this vision to avoid ending up with a European version of a surveillance-based, oligarchic business model?
The North Star that Europe should be following is not sovereignty for sovereignty’s sake, but the purpose of achieving that sovereignty. Europe, or at least most of it, is one of the last bastions of democracy in the world. It upholds human rights and freedoms, and it needs tech sovereignty to help advance these values. As long as it remains extremely dependent on systems being developed in a very authoritarian way and seeking to undermine European societies, it cannot uphold its values. Sovereignty as a concept tends to become vague if you just start thinking about it as a goal in and of itself. It should be a means towards a purpose.
What kind of coalitions, constituencies, or sites of conflict should we be starting from in order to build that alternative vision?
First of all, we should understand AI governance in an expansive way. Strengthening labour rights and protecting civil society, for example, counts as AI governance. Students, teachers, and health care workers should also be at the table to help shape and contest how AI is developed and deployed in society. Education is key, both as a sector and a coalition that we need to bring in, and in terms of public education around AI.
Another important aspect is funding AI researchers and entrepreneurs to figure out the more technical aspects of what you are trying to build and how you want to apply it. But you can’t just give them money without allowing the rest of society a say; otherwise, you are just replicating the problems we have in the US, where it’s essentially technocrats or techno-authoritarians calling the shots.
You are not starting from scratch. A lot of work has already been done to establish the common goals society should strive towards. I’m talking about the Sustainable Development Goals, for example. That could be a starting point for figuring out how to fund AI innovation.
The narratives around current AI technologies are extremely polarised. On one side, you have those who say there’s nothing particularly new about these technologies, and that the bubble is about to burst. On the other, you have those who believe AI will transform our lives, revolutionise supply chains, and make us all redundant. What do you make of these different stances?
It’s true, many people tend to exaggerate the benefits without considering the costs of these technologies, or overemphasise the costs without seeing the benefits. But I think there’s quite a large percentage of people who agree that we want the benefits of AI, but not at the cost of our own ability to live a dignified life, with access to good health care, economic opportunity, a healthy environment, and a flourishing planet. That’s the most balanced approach to take.
AI is a name for many different types of technologies, and so we should be thinking carefully about which ones we invest in, which new ones we want to build, because they each come with a different set of trade-offs; some of those align with the future that we want to build, and some others don’t. We should choose wisely.
What’s your own relationship with existing AI tools?
I don’t use them for three main reasons. One, the ethical stance. Two, the privacy stance: I’m investigating these companies, and they are ultimately the most effective surveillance technologies ever invented. Three, these tools are not very useful for the particular skill set that I need to perform my job.
I’m not saying that everyone should do the same. I’ve talked with many people who clearly get something out of these tools, or are required to use them as part of their professional role. Rather, the way to get out of the current problem is to do the same that we’ve done with other industries in the past, like the fashion industry. When we discovered that many of our clothes were being produced in a really horrific way, we didn’t just get rid of clothes. There was a lot of consumer and worker organising, government regulation, and general public backlash that reined in the industry. Of course, the fashion industry is still far from perfect, but at least there are alternatives now, and consumers can vote with their wallets over what kind of supply chain and corporate practices they endorse.
We need to do the same with AI. There are not enough alternatives for users right now, and for that to change, there needs to be more labour organising and more public backlash. Those who can abstain from these technologies should absolutely boycott them, and governments should step in to help create new markets for alternatives. Consumers have an important lever to pull, but it’s not the only one. As people, as individuals, we also wear the hats of the worker, the educator, and the citizen, and we should think about how to use all those roles in relation to AI development to make it better.
BRAVE NEW EUROPE is one of the very few Resistance Media in Europe. We publish expert analyses and reports by some of the leading thinkers from across the world who you will not find in state and corporate mainstream media. Support us in our work
To donate please go HERE .
The post Toppling the AI Empire: Interview with Karen Hao appeared first on Brave New Europe .