By the Marx Memorial Library

April 8, 2024  Morning Star, “Full Marx” feature

 

The impact of AI depends on who controls it and for what purpose, argues the Marx Memorial Library.

Popular  debate over the implications of artificial intelligence (AI) tends to be couched in terms of utopias or dystopias.

Will AI lead to some benign post-capitalist paradise — “fully automated luxury communism” perhaps, or to a hellish “technofeudalism”? The answer of course is probably neither (though, yes, capitalism could get even worse!).

Policy debate tends by contrast to focus on responses — education, re-skilling, the need to remain competitive by removing barriers to the adoption of AI, and more vague exhortations that “we” need to make sure that the potential benefits are realised. (Sometimes a virtuous “and shared” is added).

[British Prime Minister] Rishi Sunak’s much-hyped AI Summit in November 2023 at Bletchley Park (once the top-secret home of the World War II code-breakers) was a cosy natter between the 28 country representatives and the world’s major tech corporations — with SpaceX CEO Elon Musk prominently featured. It produced nothing but a vague joint communiqué warning of the risks from so-called “frontier” AI models and calling for “inclusive global dialogue” plus an (entirely voluntary) agreement on safety testing.

But technology is not autonomous. Within capitalism it is introduced for profit or control — usually both. In computing, as in the history of technology generally, the replacement, routinisation and regulation of human labour in search of profit has been a prime driver of innovation.

In the Communist Manifesto Marx and Engels emphasised the dynamism of technology under capitalism. Constant “revolutionising [of] the instruments of production,” they wrote, leads to changes in the nature of production, destabilising everything.

In his manuscripts written in preparation for what would become his Contribution to the Critique of Political Economy (published in 1859) and Capital (1867) Marx makes the prophetic remark that within capitalism, technology would culminate in an

“automatic system of machinery […] set in motion by an automaton, a moving power that moves itself this automaton consisting of numerous mechanical and intellectual organs, so that the workers themselves are cast merely as its conscious linkages.”

Marx of course had never encountered a computer, but he was aware of — and influenced by — the pioneering work of Charles Babbage, who, with Ada Lovelace (arguably the first “computer programmer”), were “inventors” of the difference engine (a calculator of mechanical tables) and the analytical engine, essentially the first-ever digital computer, the first full-sized working version of which was only completed in 2002 (it’s on show in London’s Science Museum). Both extended the data-encoding principles of the Jacquard loom (which used punched card to program complex patterns on woven cloth) to the automation of intellectual tasks.

Marx was writing primarily of physical machinery producing, in factories (with a little help from the “hands” of course), commodities that had previously been produced literally “by hand,” often in cottage industries. But his analysis applies equally to information processing.

As the former governor of the Bank of England, Mark Carney observed: “If you substitute platforms for textile mills, machine learning for steam engines, Twitter for the telegraphy, you have exactly the same dynamics as existed 150 years ago, when Karl Marx was scribbling the Communist Manifesto in the reading room of the British Libraries.”

Carney of course is no Marxist but he and others clearly understand something of what Marx and Engels wrote about automation and the contradictions it manifests within capitalism.

Most recently rapid increases in the speed of computation have led to the emergence of what is known as “deep learning” — “based on large networks of artificial neurons,
designed to emulate most basic functions of biological neural intelligence” – in which high-level symbolic or representational structure is learned from raw data. Examples range from facial recognition systems and Google’s Deep Mind which can “learn” to play videogames based only on the pixel values of a computer-screen, to “big data” in healthcare and automated high-frequency trading.

One system currently in the news is ChatGPT, a “chat-bot” developed by OpenAI and launched in November 2022. ChatGPT (GPT stands for generative pre-trained transformer) interacts with users in a conversational way through a dialogue which will respond to questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests. ChatGPT and its competitor, Google Bard are perhaps the closest that AI has got to date to meeting the “Turing test” — automated behaviour which cannot (easily) be distinguished from that of a  human being.

More prosaically AI has made possible a new “digital Taylorism” — “scientific management” which now operates through the whole supply chain, from production to consumption and it affects everyone.

As Marx insisted, there is nothing “inevitable” about the applications of technology or their consequences. All involve choices and wherever there is a choice there are alternatives. In all this of course, the concept of a fourth industrial revolution distances the word “revolution” from its essential Marxist meaning (the transfer- ence of state power from one class to another).

In this regard, Marx saw technology, within capitalism, not just as a means of increas- ing profits but also about control – control over the worker and, by displacing the worker, controlling the work process directly. Marx declared: “Machinery does not just act as a superior competitor to the worker, always on the point of making him superfluous. It is a power inimical to him, and capital proclaims this fact loudly and deliberately, as well as making use of it.

Applications today range from robot welding and ploughing through supermarket self-service checkouts and online ordering to journalism, accountancy and financial services. For Marx, technology was “the most powerful weapon for suppressing strikes, those periodic revolts of the working class against the autocracy of capital.”

Today, technology is also about controlling the consumer, from credit cards to predictive advertising based on past buying habits. And it is also about a wider programme of social control. The reality is that at the forefront of development of AI are new technologies of surveillance, repression — and war.

What some of the current debates over the latest developments in AI have done is to highlight the contrasts between what is and what could be. Marx argued that within socialism, technology would “redound to the benefit of emancipated labour, and is the condition of its emancipation.”

Humans, once freed from the bonds of soul-crushing capitalist labour, would develop new means of social thought and cooperation outside of the wage relation that frames most of our interactions under capitalism.

Increasingly the questions are asked: given the immense potential of AI to replace boring, repetitive and dangerous work, and to make previously specialised intellectual and physical tasks more accessible, why is it that leisure has not grown and the boundaries between it and “work” are as hard as ever?

Why, despite the tremendous increases in productivity has the pension age been increased? Why is inequality and poverty increasing? Why does the liberatory potential of IT contrast with the reality of covert surveillance and the harvesting of data for profit by Google, Facebook and the like, threatening a dystopia of control?

The reason is, of course, because AI development is almost exclusively in the hands of capital. As a recent editorial in this paper declared, the ongoing Post Office/Fujitsu/ Horizon scandal (the most recently visible tip of the AI iceberg) “sums up the horrors of class rule.”

The real threat of AI and machine learning today is the consolidation of corporate power. As the second part of this Q&A will argue, the challenge to the left is to develop appropriate strategies to meet this threat.

 

The Marx Memorial Library’s rich spring programme continues on April 10 and 18 with events celebrating the 50th anniversary of the Portuguese revolution, and on Saturday April 13 there’s one of MML’s famous book sales. Details on the Marx Memorial Library website www.marx-memorial-library.org.uk.