Post-Work: Nick Srnicek interviewed about A.I. as a social good - and why your dishwasher has an agenda


A very interesting interview [PDF here] in Renewal magazine with Nick Srnicek, co-author with Alex Williams of Inventing the Future: Postcapitalism and a World Without Work.

Inventing the Future, along with Paul Mason's Postcapitalism book, has been considerably influential in some of the wider discussions around the economics of the Corbyn project. Though it would be safe to say that no major Labour politician (and certainly not the Shadow Chancellor John McDonnell) has been as radical as these writers.

They call for shorter working weeks and basic income, as a response to the coming waves of automation of existing jobs. The politicians still seem trapped in a discourse of "jobs-first Brexit" and "improved productivity". 

This Renewal interview is interesting for us because it grapples with many of the bigger issues about popular control of our futures that we want to explore in our A/UK method and practices.

For example, Srnicek alerts us to the way that general artificial intelligence could become the property of just a few companies: 

What’s really interesting to me is in the past year or two, Google has started announcing itself in interviews as an AI-first company. It no longer positions itself as a search engine company. Likewise, Amazon is turning into an AI-first company. Facebook is turning into an AI-first company. Alibaba is turning into an AI-first company. There’s a sort of convergence amongst all these major platforms, towards AI being their dominant focus.

I think this is for a couple of reasons. One is that only these major companies can do AI. The way that we do AI today, through machine learning, requires a lot of data. So, you have to throw a lot of data into these algorithms, train them, and then they can do these tasks. That means only companies with a lot of data can do it, which happens to be these major platforms that have been collecting all this data. So a handful of companies that can do AI are now competing against each other to have control over artificial intelligence, which allows them to improve their services and gain competitive advantage that way.

On current trends, I think AI will become something that will be owned by a few companies, and will be rented out to everyone else as a fee-based service. Every other company in the world will need it: it will be crucial to survive in the future of capitalism. But it will be dominated by one or two suppliers, Amazon and Google most likely (at least in the Western world)....

Nick goes on to suggest how we might think through how we "own" our own data, rather than surrender it to these megacorps:

I’ll give an example: something like health data. You might train an algorithm on health data to be able to pick up, say, the likelihood of a tumour that’s cancerous or not. Clearly a useful application of these technologies. If the ownership of that data is with the public, it should change the ownership status of any algorithm or app or service derived from the data.

Rather than thinking about Google, for instance, as owning the data, the hardware/software to train algorithms, and the eventual user-facing products built from them – we might instead conceive of Google’s role as something more akin to an algorithmic factory. Their training of an algorithm is a process which turns raw publicly-owned data into something which is more readily usable.

[Google] could easily be contracted and remunerated for performing this work on the original data, but with the proviso that the end product would have ownership maintained by the public whose data it is derived from. That’s one way to push back against these companies that have control of it all. 

Srnicek's coming book is called After Work, and he previews some of it by talking about how we don't reap the time benefits of new domestic technology - because we use it to expend labour improving the maintenance standards of our lives:

There is this interesting tendency in the history of domestic technologies where something like the dishwasher comes in, and you’d think it’s going to reduce the amount of work that gets done, but what actually happens is that the standards of cleanliness just go up and up and up. And so it doesn’t actually reduce work, it just means that the standards get higher, and you have to keep dishes and clothes and the house cleaner than you ever had to beforehand.

Over the course of the twentieth century, the amount of housework done per person has barely budged, despite all the new technologies introduced. This ratcheting up of standards tends to subtly intervene everywhere. Even when you look at a lot of post-work writing, oftentimes they’ll talk about how glorious it will be that we can make these fantastic meals, once we don’t have to worry about our jobs anymore.

We should recognise that this is a subtle re-imposition of work: that we need to make fantastic meals for all our friends, which can be great for some people who love cooking, but a nightmare for those who don’t. We have to be wary about this re-imposition of higher and higher standards which compels us to do more and more social work.


Childcare is a really good example as well. There’s an interesting phenomenon in recent decades where parents are doing more and more waged work, but childcare hours are also going up – precisely the opposite of what you would expect. And that’s because the competitive demands of contemporary childcare are forcing parents to ever more tightly manage their children’s lives in some desperate attempt to optimise the perfect child.

We need to be aware of these tendencies (and in this case, their market-driven origins) and we need to think about how to manage it in some sense.

Our main beef with Srnicek is his typification of "folk politics" - horizontal, local, intimate, feelings-and-authenticity led - as being less effective than a bigger, more vertical, more planned and state-level shaping of society. We think that the "folk" level can be much more ambitious about directing bigger systems, even if only by example, than this. It shouldn't be an either/or. 

 As Indra Adnan wrote in a recent Alternative Editorial on empowered communities:

We need sustainable, resilient yet fluid units of operation. Less overbearing superpowers (or even large nations) which have no possibility of meaningful connection to the citizens. And more small nations, or more regions, cities, towns, even villages, where communities can experience autonomy while still feeling the relationship and belonging that is essential to human flourishing.

These communities, whilst connecting internally, also have to network with other communities actively. They must resist the pull-back to the hierarchical structures of the past, inventing new forms of governance to service the future they are collectively imagining.