ImperialAI: the empire strikes back

An essential conversation with journalist and author, Karen Hao, about OpenAI and its ambition to dominate and conquer all before it

A note on what I’m doing and why. I’m an investigative journalist who worked for the Guardian for 20 years latterly investigating the intersection of politics and technology that included 2018’s exposé of the Cambridge Analytica/Facebook scandal. The opaque and unaccountable Silicon Valley companies that facilitated both Brexit and Trump are now key players in an accelerating global axis of autocracy. I believe this is a new form and type of power that I’m committed to keep on exposing: Broligarchy.

I’m posting a conversation that I had earlier with Karen Hao, an ex-WSJ journalist who was the first reporter to profile OpenAI and whose book, Empire of AI, has flown into the bestseller charts.

But first enjoy this spectacular piece of AI art that I just made with OpenAI’s image generator. The prompt was ‘Darth Vader but with the face of Sam Altman’. As you can see, this worldbeating technology company is exactly who we should be entrusting our critical infrastructure to.

Consider that a scene setter for a conversation which was a rebuttal of all the main talking points and myths about OpenAI. Karen was the first journalist to gain access to the company and profile it in this piece for MIT Technology Review. She persuaded the company to allow her to embed herself in their offices and talk freely to staff. This was back in the days that OpenAI was purporting to be a non-profit dedicated to saving the world from dangerous out-of-control AI (barely five years ago).

As she explains in the conversation, Altman was shocked at the disconnect between his vision of what the company was doing and what she found through her reporting (and led to her being iced out from any communication with the company). And it’s that disconnect that she goes on to explore in her book.

In particular, it was the culture of secrecy at the company that made her suspicious about the company’s true nature and that curiosity and scepticism led to her first prescient piece and served her well in the ensuing book.

It was a fascinating conversation with someone who has studied, in detail, what is currently one of the most influential companies on the planet and who has thought long and hard about what this company is, where it is going, and what it means for the rest of us.(There’s a transcript here if you prefer to read.)

What I so appreciate about Karen’s take on this company is how she has stood back from the vast majority of beat reporting on this subject. So much of that is rooted in and influenced by the tropes and preoccupations of business journalism. That’s mostly focussed on the (admittedly astronomic) numbers, the millions poured into the company, the Byzantine boardroom plots, the business case and the rift between the founders, Sam Altman and Elon Musk, that makes teen girl dramas look mature and restrained.

These narratives largely echoe the company’s own talking points: that AI will either save the world, or destroy it, but either way it’s unstoppable.

What Karen does is look at it through a longer lens and sees a direct parallel between its messianic mission - which she describes as ‘cult-like’ - and previous historical projects, namely that of the great empire builders. And in this version of what’s happening, we are all subjects, enslaved people, pawns in a vast, rapacious power grab.

These are two vastly different stories and there is no doubt which of these narratives is winning. A point that Karen brilliantly expanded:

“I think it's absolutely right that the storytelling is so critical to understanding how these companies operate. And ultimately, you know, the argument that I make is that we really need to think of these companies as new forms of empire. And a pillar of empire building is the narrative that is wrapped around what they're doing and the consequences of it. You know, historically, empires always had this banner of we are ultimately engaging in all this plundering and labour exploitation and aggressive expansion in the name of progress, LIke that was a really critical part of justifying what they did enabling widespread public support within the british empire at the time or within other empires at the time.

“And that's essentially what is happening in Silicon Valley, that they're using progress and this imperative to move forward and his very narrow definition of what progress is as specifically as like technical advancement or advancement of AI systems to portray themselves and the quests and why they need all of these resources. And to your point that there's kind of like this really interesting dynamic of... Not only do they have a narrative of the utopia, they also have this counter narrative of this dystopia.

“And ultimately, it's kind of just the same thing.”

There were so many interesting details in the conversation. From the idea that ChatGPT is a reflection of Altman’s own personality. Karen portrays him as a people pleaser, one of the most successful fundraisers in Silicon Valley history, someone who manages to be all things to all people, reflecting back to them a version of us to ourselves that we want to believe in.

“There is this really striking similarity between the way that Altman is often accused of saying what people need to hear and the way that ChatGPT just spews and mirrors back what people need to hear. And of course, we're now seeing some really intense psychological effects of that, where people become so deeply attached, enamored to this... And that's the way that people feel—like, people feel that strongly about Sam Altman, too.

“They either believe that he's, you know, the next messiah, or they vehemently hate him. And I do think that it's kind of—the root cause of people's feelings towards Altman and people's feelings towards ChatGPTt is quite similar.”

And this is what she has to say about the Memorandum of Understanding that the UK government signed last week with OpenAI, It’s an act, I believe, of profound, inexcusable utterly naive, national self-harm. I wish that was hyperbole, but here’s what Karen has to say:

“It's so alarming because, you know, going back to this idea that these companies are empires, ultimately the only stopgap measure that we really have at the moment, the only arguably higher power than these companies is the US and UK governments.

“There really aren't that many other governments around the world that could regulate these companies if they wanted to. And now OpenAI and all of these other companies are striking deals with the exact entities that should actually be holding these companies accountable. And they're replacing essential services with AI models, or at least trying to.

“And that not only means that these governments are opening up citizen data in order to train these models to then attempt to replace these services, they're also laying off civil workers, and they're also trying potentially going to get completely false information because part of the feature of these models is that they are not accurate 100% of the time.

|And so they're basically just co-opting the government, the one entity that could be counterbalanced to them, while also undermining and gouging out the role that government has been able to play for the public and society building at large.

“I finished writing the book before any of this happened, before the Trump administration came into power, and it has been incredibly alarming to just see how this frame of understanding these companies as Empire has just come to be more and more true by the day.

“Because they have, at this point, consolidated an extraordinary amount of political and economic power to the point where they're really the apex predator now. There really is no other entity that is sufficiently capable anymore of holding them accountable and forcing them to abide by laws.”

She’s also as scathing as I am about the deals that news organisations are striking with AI companies. Why would any news organisation be in a ‘strategic partnership’ with a company whose actions need to scrutinised not licensed? It’s also, she argues, hugely shortsighted, a one-time act that can never be undone and that previous deals with tech companies shows will last for only as long as it serves the purpose of these companies.

Karen does bring the hope, however. In the book, she reported on communities who are fighting back including a grassroots successful effort in Chile. The answer, she says, will come from not from politicians or governments, but from the ground up.

Anyway, I do recommend listening to the whole thing if you want a take on this technology that’s counter to most of what you’ll read in the mainstream press. Thanks to those who tuned in and to Manasa Narayanyan at the Citizens for organising. If you’re interested in being part of a community striving to hold these companies accountable, do sign up to its Substack to be kept updated. If you have any questions for Karen, do post them in the comments and I’ll ask her if she can hop in and answer them.