I enjoyed yesterday. I took the day very slowly. Jacqueline and I went to the Welney Wetland and Wildlife Trust reserve, a frequent destination for us, and took a walk around the south end of the extensive reserve. We met just one other person when doing so. It was intensely relaxing, and time in nature is always good for the soul.
I deliberately tried not to write very much during the course of the day, but I spent the afternoon letting my thoughts wander, as it was too warm to even head out for a coffee.
I reflected on the nature of money and realised that I do need to write more about this because I still do not think that we have an adequate definition of money as yet, despite the work of the likes of Christine Desan and the late David Graeber.
I also reflected on Marx's labour theory of value and conversations I had on Friday with colleagues at Sheffield on James Meadway's strange, and profoundly neoliberal, interpretation of this theory, as a result of which he thinks that no labour is of any value at all unless it is sold in a marketplace where capitalist exploitation of it can take place. Again, I need to write more about this.
Then I reflected on AI, and two things jumped out at me.
One was a comment made by Paul Krugman in a podcast he published, I think yesterday, with Heather Cox Richardson, which was recommended in a comment on this blog. His observation was that in previous bubbles, like the British canal boom of 1800, the British railway boom of the 1840s, the American railway boom of the 1860s and 1870s, the US car company-driven boom of the 1920s and the dot-com boom of the 1990s, whilst there always was, eventually, a bust, the funds that had reached companies for investment did, in each case, leave physical infrastructure that proved to be of long-term benefit.
The canals lasted in industrial use for 150 years. Many of the railway lines built during these booms are still in use. Cars have developed further, but they became a mass means of transport as a result of investment in the 1920s. The legacy of the dot-com boom was not the technology created for that platform at that time, but the investment in fibre cabling, which has taken 20 years to realise its potential and now drives broadband.
His concern, and I think it was a reasonable one, was that there is no evidence at all that what is now happening as a result of the current AI boom will create anything of lasting value. We do not know whether the hardware being built in the form of data centres will prove sustainable. Already, 20% of all electricity in Ireland is being used to power data centres. This makes no sense, and replication on this scale appears impossible.
Likewise, when we look at the software and related IT, the potential rate of advance in technology in this sector is so great, if it lives up to its promise, that everything being created now will be of little use in only two or three years' time.
His point was straightforward. When this crash happens, as we know it will, the cost will be almost entirely financial, and there will be no investment legacy to enjoy.
Then Paul Krugman and Heather Cox Richardson made points curiously echoed by Camilla Cavendish, in the Financial Times, which I also read yesterday.
Everywhere, it seems, people are resenting AI. They hate the way it is disrupting job markets. They hate the uncertainty it is creating. They think it is going too fast. They think it is concentrating wealth in a dangerous fashion. They believe it will destabilise society. They cannot understand where this might end.
This is not the sentiment of older generations toward new technologies, as is usually found. It is, in fact, most especially heard amongst the young, who have been exposed to this technology and can see its consequences for them, most of which are adverse.
In that context, Camilla Cavendish, an independent peer appointed by David Cameron for her services to him as a political adviser, posed what I think might be a very realistic question. She asked when we might see the AI Luddites appear.
Her belief is that sometime soon we will not only see environmental protesters outside AI data centres, but also people intent on causing damage with the aim of preventing their operation because they fear the impact on their lives. She seems to think this likely, saying:
Voters won't wait forever for someone to slow things down: they may take things into their own hands. The 21st-century Luddites will be the middle classes and the working classes. And it won't be possible to shoot them all.
This is not the sort of revolutionary call to arms that is normally heard in the pages of the Financial Times, so it did, unsurprisingly, grab my attention. The obvious question to ask is whether Camilla Cavendish might be both right and wrong?
Is she right that the objections to AI that she, Paul Krugman and Heather Cox Richardson all foresee will turn to violent opposition? Is it possible that the world's people might, when they realise the potential for some deeply unsavoury people to control their lives, object, including violently? To believe that is possible would seem, to me, to be entirely reasonable.
Might a state, believing itself to have the duty to uphold supposed market power, in that situation attempt to shoot people, as Camilla Cavendish obviously thinks possible? Again, I think it is entirely reasonable to think so.
Will it, however, be unable to shoot them all? That ultimately is the big question because all conflict is eventually won by attrition. Either, as in the case of the Manchester mill owners who bought off Ned Ludd and his opponents, wealth is used to buy submission, or those objecting win by sheer weight of numbers, whatever the cost to some along the way. Tyranny, if that is where AI is headed, cannot presume that it will win in the end.
I have no answer to these questions. But the risks that they seek to appraise are, without doubt, real.
I both like the potential of AI to transform some necessary tasks in society and to free people from drudgery, if that is its potential. That we could use AI to create better lives on the basis of shared wealth is, I think, indisputable.
That much of this industry does, however, appear to be controlled by people who have no such objective also appears to be true. Is it quite reasonable, as a result, to live in fear of where their ambitions might lead us?
That is what I spent the end of my day thinking about, and this morning I still felt those thoughts worth sharing.
What do you think?
Thanks for reading this post.
You can share this post on social media of your choice by clicking these icons:
There are links to this blog's glossary in the above post that explain technical terms used in it. Follow them for more explanations.
You can subscribe to this blog's daily email here.
And if you would like to support this blog you can, here:
Buy me a coffee!
Remember my 10 Commandments for AI? I sent them in a letter to the FT. They didn’t publish it. I sent them to my MP, Anna Sabine. No reply. Pity. Here they are again, for old time’s sake. I think they might have been useful.
1. AI shall serve humanity, not govern it.
2. Human authority shall prevail over machine output.
3. AI shall not decide matters of rights or liberty without human judgement.
4. Responsibility for AI use shall rest with humans, never machines.
5. People shall be told when AI affects decisions about their lives.
6. Everyone shall have the right to human review and appeal.
7. AI shall not deceive, manipulate or impersonate.
8. AI shall not enable unjust surveillance or profiling.
9. The public good shall prevail over profit or power.
10. AI shall remain a tool, never an authority.
I amy use these on the blog. Thank you.
Don’t forget Isaac Asimov’s three laws of robotics from 1942 (robotics is just one application of AI).
https://en.wikipedia.org/wiki/Three_Laws_of_Robotics
I think that AI is a useful tool that arrived at a bad time for society. Things are bad for society because having been told for 40-50 years now that Government monopoly is bad (which was a lie anyway, it was always a public/private set up) we are now replacing it with private monopolies which are less accountable. It is this monopolistic attitude that empowers AI potential so darkly, an attitude seen in markets and recognised by the ancient Chinese, Frederic Bastiat and Adam Smith alike but conveniently ignored by North American capitalism. Government has more or less given away its powers to private corporations. Neo-liberalism, Thatcherism, Reaganism – whatever one calls it – was not about balancing both public and private, it was about creating private monopoly power. It was funded by such power when it was the Mont Pelerin Society.
Most AI I use even at work is based on extractive principles: Sharepoint tells you what everyone else is looking at when I’m frankly not interested in what everyone else is looking at; Adobe pops up wanting to sell me enhanced versions of itself when I’m looking at complicated drawings and makes me have to interrupt my workflow. Microsoft made me buy a new PC at home because they stopped supporting an older version (built in obsolescence in electronics – really?); Amazon pumps up prices on items that are short on stock; cars are made these days that are almost impossible to work at home on that compel you to go back to expensive dealerships to get fixed, monopolising your ownership of a manufacturer’s product. My feeling is that AI’s potential to aid and abet the human proclivity for monopoly power is therefore axiomatic.
Much to agree with
I agree, A.I. as a tool – opening the question who uses it – as a tool?. You listed negative uses (fair enough).
But there are positive ones, certainly in the area of number crunching. Other area( a short digression): furniture. I make furniture – with a focus on bringing out the beauty of the wood and to make something that will last centuries. There is no reason why everybody could not have several beautiful peices of furniture – custom made & lasting centuries (& priced fairly). The hard bit is the design and wood selection (& the wood is not even expensive – just bought half a tonne of cherry boards for £140). Once trained, A.I. is perfect for design and wood selection (imaging). Why should it be only the rich that enjoy beautiful furniture? There is also, as we can see with the likes of Palantir, evil applications (societal control). It is then down to the politicos to do something (ref London Mayor, the Metro Plods and cancelling the contract with Palantir). I’m not happy with the price explosion of I.T shares, & the (out of control) construction of data centres (a function of exploding IT share prices) the data centre growth needs political control – sadly lacking in Ireland and many other countries, where politicos have “bought the lie”.
Thank you
I might point of of course that if it is to shoot protesters the state needs people to do it.
Who may well share the same fears about AI and of course might not be as enamoured of being in its service as they once were – look at the Velvet Revolution
Agreed
A.I. controlled drones – deployed in Ukraine as I write.
“Battlefield-tested (in Gaza/Ukraine) as they boast at the UK arms fairs, when selling them.
There has been lots of discussion about semi-autonomous drones in articles following their use by the IDF and the target selection software that guides them and does the calculations on how much “collateral damage” (low value human? capital) is permissible depending on the seniority and importance of the intended target.
But when it comes to snipers shooting children in the head, they don’t need AI. Hind Rajab and the ambulances sent to save her weren’t destroyed by AI but by living breathing soldiers suffering from complete moral collapse.
Have you seen the robot dogs being tested?
I went down to Kingswear by train from Westbury, along the cliffs at Dawlish and through the Dart Valley.
Then boarded Waverley for a sail to Torquay and back along the Devon coast, she went up the Dart as far as Greenway to turn on her return.
But I had a fascinating conversation about the economics of steamer operation on the South Coast and how the P&A Campbell model of large fast ships racing along the coast didnt work from Torquay……………
I’ll just follow up my last post with what I’ve just seen happen on Amazon.
I was looking for a book on the humble British Rail diesel shunter, and went on Amazon to see if it was there (other outlets did not have it at all). One copy was, second hand for under £10 and I could get it for free if I joined Amazon prime. So I made a note of it and did something else in the mean time. Going back the next day, it was now over double the price! So much for Amazon and its ‘buying power’ bringing you great deals.
So what happened then? Did AI sense interest, reviewed market availability in a few nano seconds and decided to cash in? It was not a seller affiliated in the Amazon market place, it was for sale by Amazon. So I walked away. But as you can tell, it made me think. And it’s got me worried too about how markets behave in the face of scarcity. This was over a cheaply produced book about a shunting locomotives and they even wanted to exploit that!! And these outlets will selling us stuff that might become even more scarce because of Trumps aggression in Iran.
Hmmmmm………………………
I use Abebooks
Yes I know it is Amazon owned but it supports small book dealers and some are very good and prices are fair.
Ìt depends on how much tracking you permit Amazon or other marketplaces to do – these sites have been doing this via simple tracking, for many years. It can help to do your research from one device, via a VPN and with cookies & cache cleared before you start, then return to buy, on a different device, different IP address, different MAC address and a different email address. Then you MIGHT be seen as a new visitor and offered a keener price.
In the real world, haggling brings the price down. Online, the opposite happens!
A trick that makes it harder for the likes of Amazon to pull such shenanigans is either to clear your cookie cache, or browse anonymously. I’d be interested to know if you’re being quoted the same price if you browse anonymously via a VPN, so they can’t identify you.
There are undoubted benefits to AI. Fully autonomous cars will free most people from the cost and inconvenience of owning a private vehicle, at last removing the clogging mass of cars parked down both sides of the narrow streets of too many otherwise quaint English villages.
I dream of the day when a wheelchair version of Spot the Dog can provide my wife with a vehicle that can climb steps and stairs to let her into any building I can access and walk along paths and tracks perfectly smoothly without jarring and jolting her spine into unbearable agony as even the best of 4WD gophers would do at present.
Neither of these benefits will come from the giant data centres being constructed now though.
However, the most important thing that everyone must understand about AI is:
NEVER let a machine acquire the ability to defend it’s own OFF switch.
Agreed
Nor to reside in an autonomous physical body, with the ability to create more of itself.
I never underestimate the power of the state against the people. Neoliberals seem pretty good at crushing opposition, even if the result is a smoking ruin of rubble like Gaza, or a crushed S American state, like Venezuela.
I think the more likely outcome is the failure of the infrastructure. Perhaps those Irish data centres will join those derelict newly built Irish housing estates, when the Irish state can’t supply them with the electricity, cooling water and computer chips that they need so much of?
Would you invest in a data centre that relied on an English water authority such as Thames or SW Water to supply it with cooling water, or an English energy supplier and the English national and local grid to supply it with electricity?
I know nothing about AI, so I’ll stop there.
I think AI is nowhere near as useful as firms like Open.AI or Anthropic claim it is. CEOs making redundancy decisions are using AI as a cover for their own poor performance.
Yesterday I handed my notice in after an exhausting year doing the job of 3 people because I kept getting told “AI first” but the AI I had access to was nowhere near the best product on the market and even if it was, it was not going to do enough to reduce my workload by 66%.
What I actually think is happening is that CEOs are exploiting the job insecurity caused by AI to make us all work harder for less and when it goes wrong they lay people off claiming it is AI rather than their own poor performance.
I suspect that you are right. Good luck with the Hunt for another job.
I found this post fascinating but also profoundly disturbing. From what I can see there is only anecdotal evidence of the effect of AI so far. However, it does seem that the effects of AI on the real world are very mixed and many benefits seem to be marginal. Your example of the beneficial outcome of the dot.com fiasco was not predicted at the time and perhaps the same will be true of AI. I am looking forward to other people’s opinions.
Thank you Richard, and here are my thoughts for this morning –
It’s all about people:
Money is just a convenient way to settle debts. The accumulation of money and its conversion into tangible things enables greedy people to obtain power.
New things are created by people. How well things work and whether there are unintended consequences is all about the ability of the people employing the designers to foresee the future.
AI is just the latest clever tool. Like all tools, how it’s used is entirely up to the people designing it and then others using it.
Thinking is about trying to make sense of the world around us. Rich and powerful people do not want us to think. George Orwell wrote about that in 1984
Now we have rich and powerful people taking the lessons of the past and using them again. https://heatherdelaneyreese.substack.com/p/trumps-desperate-plan-to-destroy
So I am going to go for a walk in the sunshine in our beautiful countryside and look for paradise in small places…………..
I, too, went for a walk
A post today on Twitter | X
https://x.com/RnaudBertrand/status/2058450505630716013
analysing in great detail the latest Twitter | X algorithm update is illuminating. His synopsis:
It amounts “to a complete reset: your audience that you built painstakingly over years basically doesn’t matter much anymore, and it’s much – much – harder to stand out even if you’re a big account. People structurally rewarded by this algorithm are folks who: – Post visually (videos/images) – Post on globally popular topics because they clear the retrieval stage easily – Provoke strong emotional reactions – likes, replies, reposts – Don’t care about accuracy or seriousness because the algorithm doesn’t measure it – Don’t care about their existing audience because every post is judged in isolation anyway In short this new algorithm, like so many on social media, is all about maximizing whether people will engage with something – not about whether they should.”
This seems to me to be the direction of travel in social media – sensationalism and outrage.
Depressing
I have been terrified of this since realising in 2018 that BigTech had hijacked the internet and were racing for internet-enabled world domination.
Why else would they be rushing in the laws and infrastructure required for total surveillance and control? Why else would they want us to leave the ECHR? They want us completely locked down and isolated from each other in advance. They know once we twig, we won’t like it at all.
That’s why all our appeals against the draconian measures are met with the stony-cold silence of all the neoliberal governments as they betray us.
I’ve been wanting to start a new MAGA hat movement (but green). Make AI Go Away. But I only have a sphere of influence of one (my dog).
I like that
Thank you so much, insightful, and much to agree with
Thoughts you inspire are….Is the AI resistance different to canals and mill machines because its reach and energy need is voracious? Thus its impact is deeper and much more widespread?
Canals and mills were more sector specific and automation pace across sectors was more staggered, Neither did they scrape global knowledge, global resources and effectively patent it all. Thereby reduce work, ability to survive whilst increasing surveillance and social control at a global level?
There are millions of Ned’s out there to pay off.
We are ‘ghosted citizens’ we have been abandoned as you say unsurprisingly people are afraid and angry. Entirely appropriate responses matched to real existential threat posed by Billionairism (aka neoliberalism just a more relateable and accessible term outside economic realms)
Thank you
What uncanny timing for you to share this… I’ve spent thirteen years traveling around the country teaching people about the history of land rights and protest through a show I created called ‘Three Acres And A Cow’.
During the show I talk about the Swing Riots and Luddites’ motivations linked to new technology being imposed on communities in an undemocratic manner and making their lives worse… with plenty of sabotage and arson then brutal state pushback the result.
For the first time this week during shows, I have connected these historic movements with the rise of AI and suddenly that part of the show is hitting so much harder …and even got a rousing spontaneous round of applause for the first time in thirteen years.
I also read this week that in the states there are already places where electricity to data centres is being prioritised over nearby householders resulting in the breakdown of their communities.
Could AI data centres be the straw that breaks the camel of late capitalism’s back?
Maybe
And that sounds like a great show.
https://threeacresandacow.co.uk/
Ely on the 24th of October
I am speaking in Dunfermline that day. A shame.
Good point. I saw a video about electricity prices and other problems brought by a data centre in the US, on the More Perfect Union channel. There were more on the same theme.
Speaking of AI – has anyone looked at the IPO proposal for SpaceX? Available https://www.sec.gov/Archives/edgar/data/1181412/000162828026036936/spaceexplorationtechnologi.htm in initial pre-fixing date form.
This is the mission statement:
Our mission is to build the systems and technologies necessary to make life multiplanetary, to understand the true nature of the universe, and to extend the light of consciousness to the stars. To do this, we have formed the most ambitious, vertically integrated innovation engine on (and off) Earth with unmatched capabilities to rapidly manufacture and launch space-based communications that connect the world, to harness the Sun to power a truth-seeking artificial intelligence that advances scientific discovery, and ultimately to build a base on the Moon and cities on other planets
It includes the phrase “to extend the light of consciousness to the stars” 🙂
On page 11, there is a graph showing their view of the total addressable market. They suggest the bulk of the future market they will aim at is all AI based, mostly Enterprise Applications ($22.7T). Not sure what time period this is for though.
They are pitching the company as potentially worth $2T when the IPO launches.
Beware if you want to invest though – the proposal details that the shares that will be offered will only count for 1/10th of a vote compared to the shares that Musk will keep and the filing seems to suggest that only holders of the class of shares that Musk has can vote Musk off the board, and those shares can only be issued to Musk or his family… And apparently you have to sign away any rights to sue SpaceX for mismanagement if your shares lose value.
Guardian today
https://www.theguardian.com/commentisfree/2026/may/24/ai-tools-thinking-human-hard-coding-writing-technology#comment-174720379
interesting comments
It seems to me we have a choice- AI to be a useful servant giving us more time to be human OR a techno-dictatorship.
We said the same about banking. We can’t be passive about this choice, which is I think, what Richard is saying
Thank you. A good comment on a good article.
Jacqueline and I went to one of my mist frequent coffee haunts this morning, walking there and back as usual.
There were remarkably few people out, and very few people shared the beautiful walled garden we sat in for three hours. During that time we talked continually, developing ideas, exploring them in depth, and, in my case noting, them.
I came back with ideas for maybe half a dozen videos, a blog or two, and maybe a series on pedagogy. Thinking is fun.
I spent an hour with five other people sitting in silence except for one reading from Q F &P.
I let the thinking stop and allow ideas just come.
A time I have come to value.
Indeed….
I would be very interested in a series on pedagogy
It is in nascent form – but does, in other words, have a framework
AI is simply a buzzword. Technical advances have always been driven by the increasing performance, availability and reducing cost of computer hardware.
The difference now is that the AI firms are subsidising the cost of the hardware (and energy consumed) to a massive extent. It is only when this subsidising ends (and users are forced to cover the full cost) that we will see how useful the technology actually is.
The upcoming AI crash will likely trigger this reckoning.
AI data centres are very vulnerable in that they rely on two or three external inputs: electricity, water and maybe fossil fuels. Cut off either water or energy and the building is dead. Internet connectivity is also vital and somewhat vulnerable.
If hostility to data centres grows, they may become very hard to defend.
Interestingly in this week’s New Scientist leader column ” By any other name” the suggestion is that AI is snappy and more alluring than the more realistic but ” less grandiose name “machine learning””.
The models are ” machine learning ” models.
The US tech bros are all about exploitation, power, control of the populace and reducing/eradicating the costs of “lower value human capital”.
To date there is very limited evidence that the “machine learning” models are more efficient than real people.
Take the use in medicine, US research is showing that 50% of machine learning diagnosis is incorrect.
It’s still very much if you scrape garbage in you get garbage out.
[…] commentator Cliff B has offered another variation on this […]
I am not sure that the effects on working people of AI will be much different from the effects of other technological advances since time immemorial. Many people will be put out of work and employers and governments wont care much.
What worries me a great deal more is the environmental cost. I take part in a business forum where many small business owners are waxing lyrical about how much AI can help them do small, inconsequential things, and others are encouraging them to use it even more. And how often they just play with it to see what it can do, without a second’ thought about the environmental consequences of the fuel and water it uses. If AI were to charge users the environmentally adjusted cost of it, I doubt it would be used except for the most complex tasks.
Two points from this:
1 – Where I work the expectation is that AI tooling costs will jump massively pretty soon, but to give some semblance of a decent return on the capital invested, not to recognise the water and electricity costs.
2 – Originally “plastics” were manufactured from the latex collected from trees, and from potatoes. Then the oil industry developed the plastics industry to use its waste products after extracting fuel grades.
The source was almost limitless so they needed to make products which would be used in huge numbers -hence packaging….
How different our world would be if all plastics still came from potatoes and similar plant sources….
Agreed
I have yet to see any properly constructed forecast of what/who is actually going to consume the AI services provided by all these data centres. Public pushback against AI is on the rise and I wonder whether public usage that actually requires AI functionality will be limited to a small uplift on current search usage. Corporate usage will probably largely move off the cloud models onto in-house-hosted models because of the cost equation once corporates finally get around to doing the sums and balancing response speed and information reach against business need. So what is the AI actually going to be doing? It is orders of magnitude more expensive in computing resource terms to have AI do the sort of public user tasks it is being punted for in the hype than it is to use local computing facilities.
https://braddelong.substack.com/p/is-the-day-of-the-data-center-about
I have a suspicion we may end up with a very large amount of unused, but environmentally destructive and public utility resource-guzzling facilities which will then have to be repurposed or abandoned, but, by then, the damage will be done.
To me it’s very simple. There’s nothing inherently wrong with AI and its applications could benefit the whole of humanity. The problem is that AI is controlled by people who do not care about people and, at some point very soon, this needs to be radically addressed before it’s too late
Agreed
1. Is Artificial Intelligence going away? No. It is a a shiny new technology with many investors who will make fortunes even if the whole industry is a Ponzi scheme.
2. Has AI got socially valuable uses? Absolutely. Medicine. History. Economics. Any research that benefits from fast concatenation of masses of subject matter studies.
3. Is there a good way to solve the conflict over energy, water, and land? Yes. Put the data centers on satellites. It’s cold up there. They don’t need cooling water. They will use solar energy. OTOH, too much space garbage and killing darkness.
4. Can we Homo sapiens solve this problem? Can yes, political will to fight investors dubious as of now. OTOH you might have seen the judge’s rebuke to Musk’s lawyers at the recent trial over Open AI. She told Musk’s attorneys You can’t dismiss every single potential juror when you ask them their opinion of your client. Everybody hates your client. You’ll just have to suck it up.
My two euros’ worth.
Thanks
Sorry to be the reply guy, but your point 3 is seriously wrong. There is no convection in space, no conduction to other bodies. The only way to get rid of heat is radiation, which is slow. And the sun is heating one side, too.
Elon Musk disagrees.https://finance.yahoo.com/markets/article/elon-musk-wants-to-put-data-centers-in-space–heres-what-that-could-actually-look-like-143000204.html
A footnote to the discussion. Those that have read the Sci-fi novel Dune will be familiar with “mentats” – human computers.
There have been human mentats around on earth for certainly 1000 years perhaps more. The South Sea navigators, crossing 1000, 2000kms of ocean with no instruments whatsoever – apart from their brain which, through training is able to get them to the right island, spot on, time after time, by processing a vast amount of inputs. Before the last traditional one died (the obituary in the Economist is amongst the most moving I have ever read – I was awstruck at the capacity of the man to navigate) he trained others. See, humans can do it, if they want. (oh & let’s not forget – stone age people).
This is an AI generated summary from a search prompted by a memory:
Before his historic 1962 orbital spaceflight, astronaut John Glenn refused to launch until NASA’s new electronic computer calculations were verified by hand. He famously told engineers, “Get the girl to check the numbers. If she says they’re good, then I’m ready to go.”
The “girl” he referred to was legendary NASA mathematician Katherine Johnson. In an era when computers were untrusted and early machines were prone to glitches, Glenn heavily relied on Johnson’s incredible precision to ensure a safe orbit and return to Earth.
The full story of Johnson and her colleagues is told in the book and film Hidden Figures.
This one’s left me thinking all day.
Every other tech bubble left something behind that people could still use decades later. This one feels different. Data centres are out of date in three years, the software moves faster still. What’s the legacy meant to be?
What’s been on my mind is what this is doing inside organisations. I do some work in workplace mental health, and there are two completely different stories being told about AI right now. Leadership talks about productivity and freed-up time. Workers talk about exhaustion, surveillance, gaps they can’t close, and a fear that the job they trained for won’t exist in five years. Both groups are telling the truth from where they’re sitting. The question is whose truth gets believed by the people setting policy. So far it’s not the workers.
The Luddite question is interesting because I think the resistance has already started. It’s just much quieter than Camilla Cavendish’s framing. People opting out of using AI tools at work. Going off sick. Leaving sectors that have rolled it out badly. Refusing to upskill into work they don’t believe will exist. It’s not violent. It’s attrition from the other direction. And nobody’s counting it, because nobody’s looking for it.
What I keep thinking about is who controls it. The tech itself is fairly neutral. The choice of what gets automated, who carries the cost while it sorts itself out, whether the gains get shared. None of that is neutral. That’s an old question. AI hasn’t changed it.
Thank you
For those of us completely bamboozled by the technical side of AI, there remains the very old question raised by Catherine.
If technological change replaces labour with machines, then who gains and who loses, and what do we DO about that?
The sacked workers no longer have jobs, the shareholders have vastly increased profits, as do the bankers financing the new machinery.
Agricultural and textile industries provide historical case studies – the societal impacts, both domestically and internationally, are massive. It causes revolutions and wars.
The issue then becomes moral as well as practical. Should the wealth generated by these changes be shared, and what role does government have in making that happen?
Or do we just leave it to the markets, and re-run the worst aspects of the Industrial Revolution (migration, starvation, slums, begging, cholera, kids up chimneys, workhouses, plus global injustice to keep the finite planetary resouces flowing in our direction – read Dickens and Marx).
Spinning Jennies, the steam engine, the tractor, word processors, AI – is there really much difference, except in scale and pace?
Labour (provided by peasants, slaves, and an urban underclass)
Then Machines, dispensing with many manual jobs, including highly skilled ones.
Then Electronics, replacing the non-manual jobs of the middle classes. Next, quantum computing..
Where does the money go?
How is it accumulated and by whom?
Who loses out? What do they live on?
Does society/government have any role to play? (except to destroy the Unions, & suppress protest)
What will happen if we do nothing and leave it to market forces? When will the bad bits affect ME?
What will happen if we do nothing?
Which politicians are asking (let alone answering) these questions?
Very, very few.
Including far too few Greens.
I get stopped immediately by the environmental cost of the data centres. That cost is not sustainable. Is AI going to magically provide us with clean, drinkable water, once the technology has slurped up so much of the clean water we now have access to?
It’s like the nuclear power industry …it’s not clean, cheap, etc, when you take into consideration its short working life and decommissioning costs (never mind accidents.) AI data centres are just as destructive. I’m fed up with the world’s Pollyanna attitude towards new technology. I really am.
Just because we can, doesn’t mean we should.
Much to agree with
Suggest you all watch Channel 4 podcast Ways to Change the World to see the impressive Karen Hao, author of Empire of AI, a mighty900 page tome on the subject.
Regarding the use of water in cooling data centres:
The use of vast quantities of potable water for cooling is a cost cutting exercise, not a technical necessity. I used to work at a large instalation in Western Australia, built in the 1960s, where the main machinery is cooled by a closed circuit of fresh water which is itself cooled by heat exchangers using sea water. Obviously you have to build your data centre on the coast, but that is not a scarce resource, especially near the population centres here in Australia.
Alternatively, the large cooling towers used at power stations do not entirely eradicate water loss but very significantly reduce it. Remotely located data centres, powered by their own wind and solar power, cooled by huge towers or sea water are entirely possible, just more expensive.
Please note this is for information only and does not mean I’m a fan of covering the world with AI infrastructure.
This may be of interest to all: https://www.politico.eu/article/pope-leo-xiv-ai-meetings-silicon-valley-vatican/
When Leo unveils the encyclical Monday, he is expected to be joined by Christopher Olah, a co-founder of Anthropic, the American AI company that has made safety central to its public identity. Anthropic has clashed with the U.S. Defense Department over its refusal to allow its technology to be used to surveil American citizens or empower autonomous weapons, and has cultivated ties with the Vatican on AI ethics.
Thank you
If we have an E shaped society, might predatory uses of A. I result in a combination of the lower two horizotal bars which could then be enough to better manage the highest horizontal bar?
It is no revelation to state that AI is primarily a get rich quick scheme that states have invested in to get a military advantage. Almost everything else that is put out there to sell it to the world is predicated on a faster and less human-scale life. There are few – if any – uses for this technology that are truly beneficial for the world and I can’t think of anything that doesn’t already make life more difficult that would be improved by its existence.
We do not need technology to make life easier – it already requires so much more of our headspace than is healthy and militates against quiet, slow deliberation that is the basis of wisdom which already has so far to go to catch up with the tide of information that is now available.
Let us stop. The violence of taking out data centres is only necessary because those who claim to rule on our behalf have bought into this gravy train. We could, alternatively just not use it. Not buy in to the convenience that our accelerating world leaves us craving. Not get beguiled by the novelty and power of creating something that as ordinary humans we would never be able to do. This does require rejecting a part of life that is already embedded and will therefore require a serious rethink and a strategy for managing the shorter term fallout.
I am not an advocate of individualism or selfless communalism. What it requires is the realisation that we are individual expressions of a communal world and as such pushing for a life that benefits all of us using the strengths we possess is also the greatest expression of our selves.
A summary of Pope Leo’s encyclical with focus on artificial intelligence:
https://www.politico.eu/article/the-vatican-vs-killer-robots-pope-leo-xiv-takes-aim-at-ai-warfare/
The encyclical is the pope’s first since his election last year. Emphasizing its significance, the pontiff took the unusual step of presenting it in person, accompanied by Canadian tech billionaire Christopher Olah, a co-founder of AI giant Anthropic.
Leo, who holds a graduate degree in mathematics and once taught physics, does not reject artificial intelligence outright. But his encyclical emphasizes AI must remain subordinate to moral principles and new legal frameworks, ensuring new technologies “truly serve humanity.”….
“No algorithm can make war morally acceptable,” he wrote, insisting artificial intelligence must be subjected to the strictest ethical constraints and accountability in warfare. “Any technology that facilitates attacks without seeing the face of human beings lowers the moral threshold of conflict.”
I am doing my reading on this.
A video may be coming
All I have is my personal experience. “They” have cataloged my every thought to manipulate me. People who do this desire control. They want my decisions to mirror their interests. Making this more powerful will not stop.
Very interesting contribution from Pope Leo XIV today:
MAGNIFICA HUMANITAS
OF HIS HOLINESS POPE LEO XIV
ON SAFEGUARDING THE HUMAN PERSON
IN THE TIME OF ARTIFICIAL INTELLIGENCE
https://www.vatican.va/content/leo-xiv/en/encyclicals/documents/20260515-magnifica-humanitas.html
Agreed. I have been looking at it.
I do not need artificial intelligence, and I cannot see how it will improve or benefit humanity. What, exactly, is it for? It cannot invent anything, it cannot produce anything. I understand that it is fascinating to many, but that is not sufficient reason to allow it to prevail. It is terrible for the environment, which at this point should be enough to stop more data centres. It won’t, of course. Who has given permission for AI to infiltrate our lives? Did I miss the referendum? The clue is in the name- artificial. It is unnecessary, intrusive, and as far as I can see, just another method of control. I find it horrifying and inhuman. I went for a walk with a friend today to escape the heat to a lovely bird sanctuary, and found myself in tears at human idiocy. Global hunger is worse than ever, the planet is burning, but people want to talk to a robot. At times, I feel utterly powerless. This is one of those times.
I do use AI, quite a lot, Clare. Would it help if I explain how, and why? I do think it is here to stay, and the Pope’s approach is the right one. The question is about governance not existence.
Governance is an issue certainly, but of far more import is the damage data centres are doing to the environment. Richard, you are rightly worried about the impact of Iran’s actions in the Strait of Hormuz, but you seem blissfully uncaring about the far more serious threats posed by climate breakdown. However ‘helpful’ AI seems to be, its use cannot be morally justified when instead you could just employ a sub-editor or even simply reduce your output slightly.
I think that framing us too simple.
First, I am not unconcerned about climate breakdown. It is insulting for you to say so. I have written and spoken about it for decades, co-created the Green New Deal, was a professor of sustainability, and regard environmental collapse as greatest threat humanity faces.
Second, concern about one issue does not preclude concern about another. A conflict in the Middle East that disrupts energy supplies, food production and fertiliser markets can have immediate and significant consequences for millions of people. It is entirely possible to be concerned about both.
Third, the question of AI is not simply whether an individual uses it or not. The technology now exists and is being deployed at scale. The real questions are:
• what purposes it serves,
• who controls it,
• how it is regulated,
• and whether its environmental costs are justified by the benefits it provides.
I share concerns about the energy and resource demands of large data centres. Those costs need much more scrutiny than they currently receive, but your position is sanctimonious. Do you know I use more energy than you? How? The left has a nasty habit of being “holier than thou” and you are being so.
However, I am not convinced that “employ a sub-editor instead” is an adequate response. Environmental impacts should be assessed systemically, not through isolated individual choices. The issue is how we generate energy, allocate resources and regulate technologies, not simply whether one person uses a particular tool.
So I agree that AI’s environmental footprint deserves more attention. But I disagree with your suggestion that concern about climate breakdown requires the rejection of every technology that carries an environmental cost. That is not how societies make these choices, and it is not how we will solve the climate crisis.
Thank you for taking the time to reply. You are right, I can sometimes be a bit “holier than thou”, for which I apologise. I absolutely do not have that right. And it is true that if it was down to me to allocate the damage done by AI, you would be at the very front of my list for justified use. Unfortunately, using it also spreads a message that AI is OK — all nuance is lost, and that worries me. I suppose what we are debating is whether it’s possible to be a moral user of AI. That’s not a black and white question, and as a result I find myself erring on the side of caution. “You should use it, to further your important work, but nobody else should” is an uncomfortable position for me. Sometimes one has to take a moral stand, even when it’s inconvenient.
All,lorn used to be printed on paper
I am not a fan of the abuse inherent in much porn
I still used paper
We need a sophisticated and nuanced approach to AI
Noted
And thank you
AI offers great benefits for humanity in many areas, (medicine, the sciences, agriculture, environment and climate, conservation to name but a few). It would be tragic to discard those entirely because of the negatives of AI slop, commercial abuse, warmongering, surveillance and profiteering – far better to control and minimise those by legal constraint and concentrate on application of the beneficial aspects to improving human conditions.
“That we could use AI to create better lives on the basis of shared wealth is, I think, indisputable.”
Sorry but I dispute that wholeheartedly.
No matter who owns it, the computing power needed to drive even the LLMs we have today, requires enormous amounts of energy, and water.
We are already in a climate crisis, driven by over-use of fossil fuels for energy, and just as some nations are finally turning a corner to reduce fossil fuel use and transition to renewables, along comes a massive increase in energy demand.
Great news for fossil fuel companies, who are lining up to partner with data centers to power their facilities mostly by gas. While some may use or switch to, solar and batteries, this is by no means guaranteed, especially in nations with powerful fossil fuel industries.
And while our planet heats and water sources dry, the redirection of water from community use to data center use has already begun.
So, ask yourself, whose lives are going to benefit ? The already rich ? Or the majority ? Even the moderately wealthy, if they keep their jobs, are going to lose out in the long run, as people outsource more and more of their thinking to souped-up text prediction apps and lose their ability to discern not only what is true, but how they even know anything. “ChatGPT says so!”
Not to mention the tsunami of dementia that is coming.
Use your brain or lose it, folks!
I find these posts very boring. To say ChatGPT and AI has no benefits is crazy.
IF you are so pure in your electricity usage WTF are you doing here? Get off the web, now.
Please stop being stupid (and I am suggesTing you are bEing that) and understand the issue is governance, not denial, which is what you are offering.
Even I don’t claim that AI is useless. On the contrary, I think its potential is awesome. I would even (reluctantly) put up with its very dubious governance, if it had no significant environmental footprint. But the debate nobody is having is which uses do justify the significant environmental costs, and which uses definitely don’t? Of course the massive current subsidies are completely masking the real economic costs, never mind the environmental ones. I have the horrible feeling that we are blindly embracing a future in which the billionaire class will put most people out of work, then raise business costs higher than at present, with massive other impacts. Nobody has yet convinced me I am wrong to harbour such concerns.
AI is not going to go away. I have been involved with IT in one way or another since 1960s, mainly studying its applications and impact rather than its technology. Nobody accurately forecast where it is today back then. The whole world has changed in so many ways often detrimental to our environment. No doubt AI will be harmful but I am still hopeful that there will be some positives. Many forecasts predict a doomsday scenario. That seems likely to arrive with or without AI. I am sorry to have to say that I have been extremely fortunate to have lived most of my life in what will come to be regarded as a golden era.
Don’t apologise. But giving thanks is fine. I acknowledge our luck.
Living in fear of AI seems counterproductive, but ignoring it seems dangerous too. I feel like we are stuck in a weird middle ground of panic and curiosity. It is hard to know what the actual risks are versus the sci-fi horror stories. I just hope the people in charge know what they are doing, which is a terrifying thought on its own. I guess we just have to wait and see. Do you think we should be regulating this technology more strictly right now?
Yes