In Conversation with Cory Doctorow

How to describe Cory Doctorow, the polymath technology activist, artist, prizewinning, bestselling author and (since 2005) Fellow of the RSA? The man who coined the term ‘enshittification’ to describe the universal experience of how technology platforms went from ‘it’ to, well… talks life after the AI crash, who really regulates tech firms, and why we need a post-American internet. 

Stephanie Hare: You coined a term that is also the title of your most recently published book, and it became a word of the year. Can you talk us through it? 

Cory Doctorow: ‘Enshittification’ is a way of talking about how platforms decay. It describes the process of the decay, the characteristic steps of decay. It poses a theory about why the decay is taking place now, and it has a prescriptive component about policies we can make to reverse the decay and guard against it in the future.  

The decay process is: firms that are good at first to their end-users, but find a way to lock them in, and then when they know those end-users can’t leave readily, they make things worse for [them]. And the firms tempt in business customers by transferring value from users to businesses, and then those businesses become locked in, and [the firms] withdraw the value for themselves and turn into a pile of shit.  

Someone has to dare to become the disenshittification nation. They become an export powerhouse, everyone else a consumer surplus.

I can put more meat on those bones, but when you understand that pattern, and when you’ve heard a couple of examples, you can look around and go, ‘Oh yeah, this is very enshittificatory, this move by this company whose products I am regretfully dependent upon.’ 

But I think far more interesting is this thesis about why it happened, and moving it out of the squishy, neoliberal answer, which is: the almighty consumer chose unwisely and therefore, by shopping wrong with your wallet, you voted for the wrong corporations. I don’t think you can vote with your wallet. I think the reason billionaires would like you to try and vote with your wallet is they have much thicker wallets than you, and yet there’s only like… 12 of them. If we had a proper election, they would lose. But if we have a wallet election, they’re going to win.  

Our policymakers took decisions that had the absolutely foreseeable outcome of producing an enshittogenic policy environment where the worst people do the worst things they can think of and make the most money.  

Ultimately, these companies are abusing us for the same reason your dog licks its balls: because they can. No one stops them. So that’s really the most important part. 

Hare: How did we get here? 

Doctorow: Twenty-eight years ago, in 1998, Bill Clinton signed a law called the Digital Millennium Copyright Act (DMCA). Section 1201 of the Act established a new felony punishable by a five-year prison sentence and a $500,000 fine for reverse engineering and modifying technology without the manufacturer’s approval.  

Which means that if you buy a printer and it’s got some software that checks to see whether you’ve bought your ink from the same company that sold you the printer, you can’t change that software. It’s your printer. It’s your ink. [And] Parliament never passed the ‘You Must Buy HP Ink Act of 2007’. Nevertheless, it’s illegal to change your printer that belongs to you to do this perfectly lawful thing.  

And you zoom that out. You can’t change your car so that any mechanic can fix it. A farmer can’t change their tractor so that they can fix it. A hospital can’t change their ventilators so that they can fix them. You can’t change your iPhone so that you can use anyone’s app store.  

Hare: How does this affect people living outside the United States? 

Doctorow: What happened was, the US Trade Representative went to every trading partner of America and said, ‘You pass the same law to protect our tech companies from your domestic competitors so that when our tech companies abuse your population, they can’t be rescued by your technologists. In return, you will never face tariffs on your exports to America.’ 

Happy Liberation Day, right? Donald Trump has not only broken that promise [by imposing tariffs during his second administration; Editor’s note: in February 2026 the US Supreme Court ruled against Trump’s tariffs programme]. He’s also made it clear that, among the effects of having technology that is permanently tethered to an American cloud that can be modified or taken away at any time, is that all of the software you use is now part of a geopolitical coercion programme. 

Anything that’s tethered to the American cloud, like all the tractors in Europe, could be switched off at a moment’s notice, along with all the ventilators, along with all the cars, along with all the productivity software that the government and every enterprise and every household runs on.  

The only way that we can get shut of it is if we stop honouring this law. In Europe, it’s Article 6 of the Copyright Directive of 2001, which was transposed into British law, I believe, in 2003. And quite famously, European laws don’t need to be followed in the United Kingdom any more, I don’t know if you’ve heard… so long as we keep that law in our [UK] books, we cannot address that threat. As soon as we do, we could become the country that makes the tools that everyone in the world buys, that makes their technology better, even if those tools are illegal in that country, because how are you going to stop the export?  

Hare: Is this interoperability? 

Doctorow: This is interoperability – making one thing work with something else. 

Hare: It also appeals to the idea of tech sovereignty.  

Doctorow: It doesn’t matter if we don’t do it, because someone else will. This is a multi-player game. It’s a game of prisoner’s dilemma with, depending on how you count it, about 200 countries in the world. There are 199 players of the prisoner’s dilemma – everyone except America. 

Someone has to dare to become the disenshittification nation! They become an export powerhouse, everyone else gets a consumer surplus.  

Hare: As we head into the so-called ‘Age of AI’, are things going to get better or worse? 

Doctorow: The most important thing about AI is not what it can and can’t do technologically. If AI were a normal technology and there wasn’t an economic bubble, we’d just call it ‘plug-ins’ and we’d say, “Oh, some of these are quite useful, some of them aren’t.” 

The unit economics of AI are so bad. People sometimes say, ‘Oh, you know, the [world wide] web lost money when it started, too.’ And it’s true, the web did lose money when it started, but every web user was cheaper to serve. You lost less money the more people were using the web, and every generation of the web was cheaper to deliver.  

AI is the reverse. Every AI user costs the AI companies more in losses. Every generation of AI is more expensive and money losing. And they don’t have a story about how the unit economics improve – their story is that we’re going to have to step up to multi-trillion dollar expenditures on training, on chips, on data centres. Under these conditions of very bad unit economics, it’s hard to know what’s going to be left when the ashes settle, so arguing about what AI can and can’t do is less important than arguing about whether there will or won’t be AI in a couple of years. 

When you think about the lie of the land after an AI crash, you’re going to have a lot of GPUs [Graphics Processing Units] at 10p on the pound. You’re going to have a lot of unemployed applied statisticians. You’re going to have a lot of open-source models that have barely been optimised [and] that will get much better if we apply even a little bit of resources to them. The things that AI can do that are useful will probably continue to happen in some form or another, but not with this kind of incredible frothy nonsense.  

You’re [also] going to have a lot of incredibly angry people, and if all you have for them politically is, ‘Sorry, the great forces of history made you unemployed’, if you do a repeat of 2009, where [governments] bail[ed] out the banks and [had] millions of evictions, you are going to pave the way for authoritarian movements of such unhinged vitriol. People are going to be so broken, especially if the only thing the Establishment has to say is, ‘This is as good as it gets.’ Anyone who says, ‘No, it can be different,’ is going to be hailed as a saviour, no matter how materially unfit they are.  

We’ve just lived through this, and not just in America, but here [in the UK] too with [Member of Parliament and leader of the Reform Party Nigel] Farage and other ‘swivel-eyed loons’ who are nevertheless credible, because the other side’s message is, ‘It can’t be any better than this.’  

If that’s all you’ve got and there is no alternative… you’re just going to drive people into the arms of fascists. 

Hare: A common argument against regulation [in the West] is: ‘Regulation will hinder innovation and then China will win.’ 

Doctorow: Well, we should hinder innovation in the ways to abuse human rights at scale!  

If you don’t do antitrust regulation, competition regulation, then what happens is all your other forms of regulation become the remit of monopolists. It’s not that the regulation goes away.  

Look at what apps can exist on an iPhone. It’s not the apps that are determined by a ministry here [in the UK] or by an EU regulation. It’s the apps that Apple will accept for the iPhone App Store. They structure that market. They also exercise editorial discretion. And a lot of people are like, ‘Well, that’s fine, I don’t want 50 fart apps in the app store.’ But that editorial discretion is incredibly politicised. 

Hare: It’s very ‘gatekeeping’. 

Doctorow: There was an app in America called ICEBlock, and it warned you if there was a man in a mask who wanted to kidnap you in your neighbourhood. And Apple said, ‘Actually, ICE agents are a protected class and apps that adversely affect them are prohibited under our terms of service.’ 

Hare: Under what grounds are they a protected class? 

Doctorow: Under the same grounds that someone is a protected class on the basis of their political beliefs or their gender. On the grounds that they don’t want tariffs on their phones!  

So [Apple] removed it, and because the device is designed so that you can’t install an app that’s not from the App Store, and because Article 6 of the Copyright Directive establishes a crime for reverse engineering an iPhone to put a different app store on it, therefore, Apple structures the market. Apple has de facto made ICEBlock illegal, not just impermissible on one platform, but illegal for everyone who owns an iPhone.  

This is a form of regulation. If it were a form of state regulation, if this was coming from a backbencher in [UK Prime Minister Keir] Starmer’s government who decided that certain apps that offended them just shouldn’t be sold, there’d be a huge hue and cry. But when a firm does it?  

When you stop regulating firms for monopoly, for competition, you lose the ability to regulate firms for all other conduct. Yet all other conduct remains regulated, but solely by that firm.  

Hare: What tool or technology does not exist right now – or is not yet good enough – and you wish it did? 

Doctorow: Universal fibre (fibre optics). It exists; we just haven’t run it everywhere. Right now, everyone’s like, ‘We should buy Elon Musk’s stupid satellite service’ which currently runs at about 1% of the speed you get from a single strand of fibre optic. But every time you add a user, it gets slower. And they’re like, ‘Well, we have to use it because we don’t know how to run wires to remote farmsteads in Wales.’ I’m sorry, we did that 120 years ago, it’s not like we’ve forgotten! Just bite the bullet, run a piece of conduit out there, fill it with a rope of fibre as thick as a baby’s arm, and deliver future-proof data forever and ever. 

Hare: What does it mean to be an effective technology activist and thinker today? 

Doctorow: You have to know what the technology can do.  

In the crypto wars – not cryptocurrency, but encryption – when we were fighting about whether there would be a ban on working encryption, our adversary said, ‘We’ll just have an encryption system that works when good guys are using it but fails when bad guys are using it.’ And we would say, ‘Wanting it badly is not enough.’ 

You do need to know what’s technologically possible. A lot of the stuff about AI, data acquisition and training, talks blithely about privacy and bias as though those were well-understood levers you could pull in model training – and they’re not. Preventing memorisation or having an AI understand which parts of its training data are private and sensitive, and which ones aren’t, and having it omit those from its outputs, or maybe even identifying them correctly and removing them from its inputs – those are not solved problems.  And [determining what is and is not infringing] is not something you can ask software to identify. 

You have to know the technology. You have to understand the economics. You have to understand the politics. You have to attune yourself, if you want to intervene in this, to something very, very important, which is administrability. How hard is it going to be to enforce this rule? That’s administrability in a nutshell: it is thinking about how you make the rule work in the world and not just about what rule you want to construct. 

Because I write science fiction novels, people sometimes think that I have views about what will happen in the future. But I actually have views about what we should do in the future, which is very different. If the future were predictable, there’d be no reason to get out of bed, right? That’s why I don’t like optimism or pessimism, because they both feel like fatalism.  

As an activist, I’m like, ‘All right, I can see one or two steps ahead, we’re going to have certain material phenomena in the world, what should we do about them as opposed to what will happen.’ Because I think what will happen is up for grabs. It’s based on what we do.