What do you think?
Rate this book
Inhuman Power: Artificial Intelligence and the Future of Capitalism
The past several years have brought staggering advances in the field of Artificial Intelligence. And Marxist analysis has to keep up: while machines were always central to Marxist analysis, modern AI is a new kind of machine that Marx could not have anticipated.
Inhuman Power explores the relationship between Marxist theory and AI through three approaches, each using the lens of a different Marxist theoretical concept. While the idea of widespread AI tends to be celebrated as much as questioned, a deeper analysis of its reach and potential produces a more complex and disturbing picture than has been identified. Inhuman Power argues that on its current trajectory, AI is likely to render humanity obsolete and that the only way to prevent it is a communist revolution.
Inhuman Power explores the relationship between Marxist theory and AI through three approaches, each using the lens of a different Marxist theoretical concept. While the idea of widespread AI tends to be celebrated as much as questioned, a deeper analysis of its reach and potential produces a more complex and disturbing picture than has been identified. Inhuman Power argues that on its current trajectory, AI is likely to render humanity obsolete and that the only way to prevent it is a communist revolution.
176 pages, Paperback
First published January 1, 2019
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 9 of 9 reviews
August 28, 2023
This is a very readable book with a lot of informative journalism. What it contributes to debates on Marxism and AI, I'm less sure. The authors align themselves with modular cognitivism (e.g. Cosmides, though their engagement with this lit is very thin) and believe innate heuristics of the kind promoted by behavioural economists explain human life activity. They even approvingly cite Yuval Harari. I can see how this can serve "analytical marxists" like Elster, but it otherwise clearly goes against Marx & Engel's own work and intentions. Since they're already of the view that general intelligence is about finding the right algorithms, it's not surprising they think any dividing lines between human and machine can be bushed aside. They are also fond of Timothy Morton and offer his argument that perhaps humans don't really have imagination, they're just running algorithms without knowing they are.
They build up to the thesis that AI can produce value just like humans. They barely engage with the vast debates on this and their engagement with Marx's value theory is weak too - Or rather, it refers to Marx‘s arguments without attaching much weight to them, seeing them as overturned by empirical findings of some general kind. As cognitive functionalists, they just think machines will produce value because they'll be doing what humans do. In fact, they will produce value better because they will not need to waste time and energy superfluous activities like breathing and eating. But it is of course precisely such superfluous activities that make value possible. Insofar as machines will need to reproduce themselves, they will not do so via the exchange relation and socially recognised labour time.
The chapter on Communism and AI basically says forget about it and since the authors don't have the best grasp of AI, there is little that is new here. The authors begin and end the book with Nick Land, concluding that we are heading for "A capitalism that happens without humans". They suggest that they depart from Land in that he suggests his vision is a good thing whereas for them it is more ambiguous.
Broad in scope and clear but shallow and disappointing.
They build up to the thesis that AI can produce value just like humans. They barely engage with the vast debates on this and their engagement with Marx's value theory is weak too - Or rather, it refers to Marx‘s arguments without attaching much weight to them, seeing them as overturned by empirical findings of some general kind. As cognitive functionalists, they just think machines will produce value because they'll be doing what humans do. In fact, they will produce value better because they will not need to waste time and energy superfluous activities like breathing and eating. But it is of course precisely such superfluous activities that make value possible. Insofar as machines will need to reproduce themselves, they will not do so via the exchange relation and socially recognised labour time.
The chapter on Communism and AI basically says forget about it and since the authors don't have the best grasp of AI, there is little that is new here. The authors begin and end the book with Nick Land, concluding that we are heading for "A capitalism that happens without humans". They suggest that they depart from Land in that he suggests his vision is a good thing whereas for them it is more ambiguous.
Broad in scope and clear but shallow and disappointing.
March 29, 2024
Muy interesante abordaje de la IA desde un lente marxista. Los autores toman los aportes del autonomismo, los aceleracionismos de izquierda y de derecha, el decrecionismo y las posturas críticas, y tratan de construir un enfoque original, que no decante ni en el utopismo ni en el ludismo.
July 4, 2021
I came to this book fairly skeptical towards its stated, and fairly dire, conclusions. But I found the authors approached the topic with much warranted skepticism themselves, and have written a serious, for the most part "realistic" analysis. Most of all the authors illustrate how well a Marxist dialectical approach to the problematics of AI works, and how AI has the potential to supercharge class warfare at the hands of capital (or, the dilution of working class power in the name of strengthening profits, if you prefer), overwhelmingly at the expense of us proles.
January 20, 2024
More than anything, the book is a speculative, relatively superficial exploration of where AI is and where it's going. The author is obviously heavily influenced by operaism/autonomism, which would be fine if he weren't trying so hard to convince the reader he isn't, and the short engagement with the kernel of Marx's analysis of capital relies on an apparently circular argument (either that or I was too stupid to follow it properly): the perfect machine is not a machine, and therefore it cannot be fixed capital and instead has to be variable capital. And AGI employed as fixed capital would be, in essence, a slave. But since a slave is a "living-labour machine" and this perfect machine isn't a machine, AGI *can't* be a slave. This confused argumentation culminates in Dyer-Witherford postulating a Westworld type scenario, the proletarianisation of AGI, and the elimination of humans from the productive process. I didn't find that particularly insightful. Admittedly, the previous 20 pages of vamping about the ontological capacity for labour (totally tangential to the value-creating, abstract labour form which defines capitalism) didn't help.
August 31, 2023
Very well done. I wish they would have expounded more on how the development of AI can lead to it being integrated into society to the point where it becomes a condition of production. There are some interesting and speculative places one could have taken that. The same goes with the latter part regarding Land’s work and the emancipation of production (from us) - an opportunity presents itself to deal with his ideas in a much more grounded sense. Fun read - easy to follow.
May 10, 2024
this was quite interesting and probs would have given it more if I had not power read it for my exams
June 23, 2026
You wouldn't think Marxist theory would provide a great lens for predicting AI's eventual effect on the global economy, but this 2019 book could surprise you. Dyer-Witheford's main question is whether AI will be a commodity (qua machines as capital investments) or go beyond that to replace humans as creators of surplus value. His answer is probably both.
If and when AI becomes AGI and conclusively assumes exclusive creative design of new products and services, while also exclusively recasting the media to condition public appetites / fetishes for its own generated offerings, it will, he argues, effectively sideline humans as meaningful economic actors (something Marx never foresaw).
Gradual or rapid expulsion from economic relevance could, per this argument, be what propels humanity on possible paths to extinction. Capitalism, he notes, is a value universe unto itself and has little patience for anything not strictly necessary to it. The fast-ev0lving AI policy generators it empowers may find little logic for maintaining universal basic incomes for great masses of extraneous non-players. Even the final handful of bloated human trillionaires may end up deported, then liquidated, as their uselessness becomes all too apparent. Dramatis finis!
If this sounds like science fiction -- and the author does in fact frequently reference SF tropes and plots -- it's because AI, especially with the threat of AGI, may in fact be science fiction coming to life. Have we perhaps reached a point where some measure of SF speculation must be allowed into our critical mix for us even to think about AI?
If and when AI becomes AGI and conclusively assumes exclusive creative design of new products and services, while also exclusively recasting the media to condition public appetites / fetishes for its own generated offerings, it will, he argues, effectively sideline humans as meaningful economic actors (something Marx never foresaw).
Gradual or rapid expulsion from economic relevance could, per this argument, be what propels humanity on possible paths to extinction. Capitalism, he notes, is a value universe unto itself and has little patience for anything not strictly necessary to it. The fast-ev0lving AI policy generators it empowers may find little logic for maintaining universal basic incomes for great masses of extraneous non-players. Even the final handful of bloated human trillionaires may end up deported, then liquidated, as their uselessness becomes all too apparent. Dramatis finis!
If this sounds like science fiction -- and the author does in fact frequently reference SF tropes and plots -- it's because AI, especially with the threat of AGI, may in fact be science fiction coming to life. Have we perhaps reached a point where some measure of SF speculation must be allowed into our critical mix for us even to think about AI?
October 16, 2024
Not sure why I did not review at the time of finishing. The single best treatment of AI through a Marxist lens, ends up in some very provocative and disturbing places. So good, I might pick it up again and re-read soon.
November 4, 2023
Perfectly readable treatise on AI and Marxism.
Displaying 1 - 9 of 9 reviews