(PART 1.)

I suggest that currently there are only three types of intelligence; the Good, the Bad and the Ugly. (Thanks to whichever Spagetti Western writer formulated that film title) The Good is Socio-biological Aware Human Intelligence (SBAHI); the Bad is Individually Applied Human Intelligence; (IAHI) and the Ugly is Artificial Intelligence (AI). I will try explain my reasoning for this analogy by what follows in the rest of this two part article. For some time now our societies have been bombarded by AI generated advertising, literature, news, music, images and information on internet blogs etc.

And the reason I classify AI as Ugly is down to five symptoms of its production and use. 1, the material used to produce the huge buildings etc., which house its production; 2, the energy sources needed to power its production processes; 3, the raw material (data) used in it’s manufacturing process; 4, the air and ground pollution caused by all its processes and 5, the social confusion its creators are causing by blurring the conceptual difference between skills and intelligence. A skill is the ability to do something complex in a sustainable and accurate manner. Skills are often impressive, but intelligence is different.

The eight characteristics (linguistic, logical, spatial, body kinesthetic, musical, interpersonal, intrapersonal)  outlined by  Huawei Technologies, in their outline of AI intelligent systems are not actually outlines of intelligence, but outlines of human socially acquired skills. However, the authors of this seminal report seem not to have realised that human intelligence is not the same as a computer processing skill, no matter how complex that computing skill is. Human intelligence is something qualitatively different.

For example, Intelligence is knowing when the exercise of any skill is accurate enough for what is needed or intended, and whether and when the exercise of a skill is appropriate or not. Speaking is a complex everyday skill but intelligence is knowing what is accurate enough to pass on to others, when it is appropriate to pass it on and when it is best not to pontificate but just to shut up. Computers and their programs, on the other hand, are not intelligent and will carry on processing output until the code corrupts, they burn out, the electricity fails or until someone switches them off. Fears about AI taking over the planet? Really? They cannot even switch themselves on.

Lots of people (and some birds) have skilful use of speech, but not everyone (or every parrot for that matter) has the intelligence to know when and how to use the words appropriately. Many people have the skill to drive or fly various vehicles at rapid speeds and yet they frequently drive too fast, too recklessly, or in hazardous conditions and die every day – often taking others with them. Their skill levels frequently override their intelligence levels. Poor James Dean. The calculation skills of AI computer algorithms is impressive, far faster and often more accurate than human calculations. The more advanced ‘expert’ systems contain knowledge bases of stored facts and inference engines which apply logical rules to the knowledge base and find patterns, that are quicker and often more accurate than a human being can do in the same time. But that is not a symptom of intelligence.

Take for example, the dogmatic certainty of the programmers of the accuracy within the UK Post Office accounting software. That reliance on machine-level processing led to hundreds of people being wrongly convicted and jailed for fraud. Some of the wrongly convicted even committed suicide in desperation or despair. When in fact it turns out that the expert AI type programs housed within the accounting system were not accurate, but faulty. The programmers exaggerated confidence in their own programming skills and their machines calculation skills over-rode the operation of their social intelligence. A human intelligence applied to that outcome (unless deeply prejudiced) would have questioned whether so many Post Office managers could be so inept or so corrupt. Not so the computer programs or their programmers. Like their computers the latter ploughed on regardless until someone eventually switched them out of the chain of misery they had created. But far too late – people had already died.

The computers or their programmers did not jail them or prompt their victims existential distress, but a chain of people who dogmatically believed in the computational accuracy of their ‘expert’ machine learning programing did. Again this example indicates that skill levels are not symptoms of intelligence even when performed by otherwise intelligent human beings. That result to my mind was another downright ugly aspect of human reliance on machine learning and its latest digital offspring Artificial Intelligence. Indeed, I would argue that reliance upon machine learning skills are already eroding the application of human intelligence and skills. If we are too tired or busy to check our spellings or calculations, we don’t need to reason why we are too busy or to tired to complain, we can just get the software programme to check both, as the Post Office elite did.  The logic being if it’s cheaper to let AI do it why pay a human intelligence to do it?

AI: Capitalism on Steroids.

Metaphorically speaking, AI is an example of capitalism on steroids. Like any large-scale-capital-funded industrial factory building in the last 200 years, AI factories are large extensive and expensive buildings, containing high cost exclusively designed machinery, large raw material inputs, high power source usage, a complex manufacturing process, and a slick but completely one-sided brand of sales promotion, to expedite its completed commodity for sale cycle of production and consumption. AI data centres and their products are no different in this capitalist motivated essence, to car factories, mechanized bakeries, or factories producing televisions, mobile phones, biscuits or just streams of decorative pound shop tat.

The output and sales are what drive the capitalist production process forward and onward, not quality, equity, safety or sustainability. In this regard, two forms of capital are normally deployed in industrial and commercial capitalist processes: Fixed capital (expended mostly on buildings, machinery etc) and Variable capital (expended mostly on wages/salaries for workers). However, with little or no new intellectual labour needed for AI data and number crunching, the variable capital required for AI and Bitcoin production salaries and wages is practically zero.

Also like previous capitalist industries, all AI outputs are being generated from processed raw materials which are repurposed and then resold as entirely new commodities. In non-digital production processes, plastics, metals and glass are the raw materials processed by workers into smart phones, cars, laptops, etc, who are employed on various pay scales to do this transformation. Hence their wages and salaries are spent in the community which helps keep the lives of the rest of those working within the capitalist mode of production afloat. In the case of AI processing the raw materials being used are not past inorganic materials produced during the planets evolution such as powders, metals, plastics or even wood, but are the recorded past examples of human literary ingenuity, mediocrity and it has to be said, even banality.

Prior to the AI digital period, a capitalist had to employ living intellectual labour in order to exploit the difference between buying it and selling its output, but that is no longer necessary for mundane intellectual production. A way now exists to put new life into the dead and archive buried intellectual labour of the past. A reversal has occurred of Marx’s vampire analogy where the owners of dead labour live by extracting material from the living. In AI the living capitalists extract intellectual material from the dead or retired intellectual workers of the past.

The raw digitised material for AI comes in the form of humanities past products of social evolution, as embedded in their words, images, and sounds, previously stored mostly on paper or digitally but now contained in massive data holding ‘clouds’ and processing computing centres. This accumulation of past, once profitable intellectual labour-power, is being re-processed and resold back to humanity, for further profit making, as something entirely new. But in fact AI is old intellectual labour of varying quality being recycled, re-packaged and resold, to make very rich people even richer.

With regard to the normal capitalist formula presented by Marx, after studying Adam Smith (and many others) to explain how capital makes profits, he identified the following sequence. A sum of money designated as Constant Capital (c) is expended on Fixed Capital forms (buildings, machines etc) and another sum expended as Circulating Capital forms (on raw and auxiliary materials of production) plus another sum expended upon Variable Capital (v) which pays contemporary labour. The workers, by being paid less value in wages and salaries, than the value they add to the commodities or services they produce, thus also produce surplus value (s). The extra unpaid labour they perform during the production process creates this surplus-value in the form extra products, which enables profits to be extracted, once the extra products are sold.

Thus Marx in Das Capital utilises an abstract capitalist investment formula of 4,000 c + 1,000 v = 5,000 to calculate the total capital invested. But since it is the new labour (v) which adds extra value to the existing commodity value once sold, the finished commodity product value under capitalist investment is intended to be as follows. 4,000 c + 1,000 v + 1,000 s = 6,000 (at a 50% surplus value rate of labour exploitation). Thus the capitaist puts 5,000 in and gets 6,000 out – in whatever form of currency the capitalist operates within. So the total Capital advanced (C) = constant capital + variable capital or (C = c+v): But the total Capital returned after sale is constant capital value + variable capital value + surplus value or (C = c+v+s); (s) being the source of profits and interest on capital. (Marx. Vol. 3, Das Capital)

The limits of surplus-value exploitation of human labour (intellectual, physical or both) in such normal capitalist cases are not just technical but physical and biological. Generally, human labour becomes exhausted after eight or ten hours of continuous value adding activity, (hence formal and informal mandated time limits for working) but for AI without a workforce there are only technical and marketing limits to the extraction and realisation of profits or super profits. So in the case of AI factory production without needing the use of current labour as an expense, the same investment calculation would be 4, 000 in + whatever the AI entrepreneur is able to charge the purchasers for whatever residual value they still think remains within global societies currently stored past intellectual labour-power output.

To the capitalist investors in AI (and of course Bitcoin) this is the same type of business opportunity that ancient and modern piracy or ancient tomb or bank raiding theft offered or still sometimes offers. They just fund the costs of getting their hands on already salable merchandise, (i.e. a pirate ship and boarding crew for piracy or a van and gang for a warehouse or bank heist. “Hey I only said blow the bloody doors off!”). And at a capitalist system level scale, therefore, this AI worker reduced production process leaves more entrepreneurial capital available to be spent on more Constant and Fixed Capital, (bigger AI processing units) or on more interest bearing Loan Capital or leaving more capital to be utilised for Financial Speculation.

AI: And Financial Speculation.

Incidentally, what AI is currently in financial terms, is a huge capital funded financial bubble in AI and Bitcoin investment. AI investment is just metaphoricaly pumping up the presssure for its price to bubble up even further and eventually burst whilst at the same time before the burst ocurrs, spilling enough spare capital elsewhere to fund other available speculative bubbles. Hence, for example, the Blair and Trump family attempts to sell future hotels and condo’s on a bulldozed and repurposed Gaza Strip, fully stripped of it’s original buildings and original population; or selling some future shares in rare earth mining consortiums located in confiscated parts of Ukraine.

We should all know by now what happens to bubbles – they burst! Just like they did spectacularly in the 1930’s and again in 2008. But because the capitalist mode of production is a socio-economic, inter-dependent social ‘system’, it is not only the bubble makers who suffer loss. Everyone in the ‘system’ is effected albeit disproportionately, but invariably it is the poorest and most vulnerable who fare the worst from a huge economic or financial bubble bursting and the inevitable slump which follows. The quotation which follows is a typical liberal, bubble-encouraging bourgeois assessment of AI. Such examples are originating from a modern version of a snake oil salesman addressing a naive and gullible public. Note the self-satisfying motive of avoiding the costs of employing paid skilled, intellectual, salaried or waged labour;

“For the first time in history, it has become realistically possible to meet the needs of the majority of the population with minimal human effort, and to provide goods, services, and knowledge abundantly, sometimes even for free, without relying on intensive wage labor or traditional bureaucratic structures” (Autumn 2025)

This promise of meeting the needs of the majority is offered with minimal effort, by the writer without mentioning that this is providing that the human need is not for a paid job in intellectual literary, graphic, or musical composition of some kind, which AI is busy replacing. Instead, AI, like previous generations of snake oil salesmen are assuring their punters, is going to cure all ills and do it for free! Is it indeed? The above idealist scenario is a slimmed down version of Elon Musk’s science fiction utopia where robots do all the manual labour and the humans get a basic income from the state to keep them alive. The basic income will supposedly come from a tax on the wealthy corporations who will clearly carry on doing all they can to avoid paying it. It’s basically a total war type imaginary economy in which every citizen consumes (food and resources) but no citizen produces – the machines do all that faster than people and the waste stuff and pollution piles up and up.

The above AI bandwagon jumpers, as with other, left, right and centre, AI advocates, make no mention of AI monopolising civic electricity and water supplies, they make no mention of the hidden social costs in job losses, no mention of accelerated environmental degradation, no mention of further pollution, climate dislocation, no mention of further species extinctions! In reality down this yellow brick road of fantasy metaphorically free meals, it is more likely that bits of our own patched together collective ingenuity, mediocrity or banality, will come back to those of us who are conned into buying the ersatz AI commodity based upon our previous collective intellectual output.

So in the case of AI, its enthusiastic promoters think we should not bother about the recycled mediocrity and occasional gems of wisdom already previously uttered, we just need to admire its sleek data-set packaged form and dream the dream of it meeting our ‘needs’. Peel back the rhetoric and the protective plastic film and plug into the fact that AI is just recycled old news, resuscitated old opinions, worn out old narratives, tattered old platitudes, long decayed old reassurances, old dreams, out of date social and scientific assumptions, clever and subtle falsehoods, past amusing anecdotes, threadbare repeated cliches and a general assortment of fabricated intellectual tat.

But the fact that such manufactured machine produced ‘brave new world’ digital culture is occuring, whilst human interactive community culture in many human communities, continues to be gutted, atrophied and undermined, doesn’t matter to the owners of capital. If the value of the last 100 plus years of recycled anthropocentric opinions and class based obsessions, can be hyped up to be of greater value than the value expended by the capitalist investor in order to recycle it, then that difference produces the profit or interest on the capital invested. The capitalist investor is only interested in the welfare and accumulation of his or her own capital not the welfare and future of citizens in general or the other particular life forms sharing the environment. The only difference will be the packaging and a new generation of slick snake oil type salesmen.

Their job will be to convince the gullible amongst us that purchasing our past collective intellectual products or crunching our accumulated statistical information (which paid workers could have done, with critically applied intelligence) will meet the needs of the majority and solve practically every known problem humanity can think of. Of course, the industrial era raw material inputs, (cottons, plastics, metals, etc) were themselves all products of stored up past human labour, which were then worked upon again by yet another set of paid workers to manufacture whatever commodities the factory or industry owner had decided to manufacture and sell. In the case of AI data centres however, there is an impossible to ignore essential difference. As already noted, the raw material in AI processing is the stored up results of past intellectual labour by living and dead authors, writers, musicians, artists, poets, publicists etc.

AI Production: Out of Sight out of Mind.

However there is another notable dimension to this digital new world. The vast AI data centres, (and their fellow travellers in making money out of selling ersatz dreams, the bitcoin number crunching centres) however are not based near locations of mass labour such as towns and cities, because these industries are not going to be crammed full of manufacturing workers, together with a sprinkling of maintenance workers. These new industrial-sized monuments of mammon are being erected in isolated places and are crammed full of banks of high powered computer hardware storage and processing units, with their own few technical and maintenance operatives.

Consequently, it has been reliably estimated that the total yearly energy consumption of bitcoin production (energy and resources used just to solve complex cryptographic puzzles, to produce a rare imaginary mathematical result) is already equal to, or more than, the power consumption of countries the size of Finland. Yes Bitcoin production is consuming more planetary resources than entire hierarchical mass society populations of modern countries! And of course, AI computer data centres are not far behind this resource consumption figure and the latest scaled up ones in the US and China are certainly in the same resource draining, polluting, eco-damaging superleague ball park as Bitcoin.

So of course, it should come as no shock that the power and water inputs for these hot running computer banks of AI and bitcoin, national and global production, are also far higher than 19th and 20th century industrial era polluting electrical and water supplies to 19th century factories and industries, which left, physical and social scars on the regions and communities in which they boomed, slumped and then went bust. These 21st century power and resource consuming capital-intensive, data processing investment factories now require their own dedicated power grids and large reservoirs of water to keep them running 24/7 and keep them optimately cool whilst doing so.

Consequently some AI providers are recommissioning abandoned decrepit nuclear power facilities, or old coal fired power stations, and even second hand large jet plane engines, patching them up to run again on full power to keep their AI banks running 24/7. If one or other of these desperate measures, is not another disaster in the making like bhopal, I will be pleased but surprised. AI is another case of out of sight, out of mind! These vast complexes are also out of hearing (noise pollution), smelling and tasting (aerosol pollution), feeling (liquid coolant pollution), and out of sight of ground saturation and detection as well as wild life micro and macro organism destruction. Yes! Doesn’t AI seem wonderful, when it avoids socio-biological focussed forms of human evaluation and is allowed to just individually evaluate itself? Like elite decisions everywhere – the dark side literally and metaphorically just ploughs on – protected and undetected – until it is too late.

Roy Ratcliffe (January 2026)

Part 2, will consider further the liberal and left perspective on AI in the period of climate change, ecological destruction, human and animal rights violations, and the comprehensive, multi dimensional biosphere pollution.