Artificial intelligence promises unprecedented prosperity. But that prosperity is becoming concentrated in remarkably few hands. A new generation of robber barons has emerged, prompting warnings of techno-feudalism: an economy in which a handful of firms own the digital infrastructure on which everyone else depends.
A defining political question of the AI age is, therefore, not whether wealth will be created, but who will own it.
The history of liberalism is, in many ways, the history of widening ownership. Nineteenth-century liberals challenged the concentration of land ownership because they understood that whoever owned the dominant asset of the age exercised disproportionate economic and political power. Today, the defining assets are no longer simply land and factories, but increasingly the intangible capital of the digital age: data, algorithms, computing power and artificial intelligence. Liberalism should respond to the AI revolution as it always has – not by resisting wealth creation, but by widening ownership.
The Liberal Democrats should answer that challenge by championing a Citizens’ Ownership Fund.
The inspiration comes from Alaska. For more than forty years, the Alaska Permanent Fund has invested part of the state’s oil wealth on behalf of every resident, paying an annual dividend while preserving the capital for future generations.
Britain should apply the same principle to the AI economy. A Citizens’ Ownership Fund would build a portfolio of shares in frontier AI and technology companies. Those investments would generate dividends paid equally to every citizen while the underlying assets remained invested. Instead of merely taxing AI after wealth has been created, Britain would own a permanent stake in creating it.
For too long governments have assumed that the public benefits from successful companies only through taxation. Liberals should think differently. Where public investment helps create extraordinary private wealth, society should sometimes receive equity rather than tax. Ownership is a better dividend than redistribution.
AI has not emerged from nowhere. It rests on decades of taxpayer-funded scientific research, world-class universities, publicly educated talent, national infrastructure and the rule of law. Where companies benefit from those public foundations, they should make a modest equity contribution to the Citizens’ Ownership Fund. This could be linked to access to national computing infrastructure or, for the largest frontier AI firms, as part of the regulatory framework governing their operation in the UK. Instead of asking society to settle only for tax receipts once wealth has been created, Britain would acquire a permanent ownership stake in the industries that public investment helped create.
Freedom is more than the absence of interference; it also means not being subject to another’s arbitrary power. Ownership provides independence. A citizen with a stake in productive capital is less dependent on employers, markets or the state alone. The dividend from a Citizens’ Ownership Fund would not make anyone wealthy, but it would give every citizen an independent stake in the nation’s prosperity. By dispersing ownership, we also disperse power.
This is not nationalisation. Companies would remain privately owned, independently managed and free to innovate. Entrepreneurs and investors would continue to prosper. However, instead of relying solely on taxing success after the event, the British people would share directly in that success through ownership.
The fund, not individual citizens, would own the shares permanently. Citizens would receive equal annual dividends, while the capital remained invested for future generations. Ownership could not gradually be bought back by the wealthy.
If AI shifts an ever greater share of national income from labour to capital, then broadening ownership of capital becomes the defining liberal challenge of the twenty-first century. As well as asking how we should regulate AI, liberals should ask who owns the wealth it creates. The answer should not be a handful of AI oligarchs. It should be every citizen.
* Daniel Duggan is a Liberal Democrat Councillor in Gateshead
Follow @libdemvoice.org on Bluesky
Like us on Facebook
Subscribe to our feed
Sign-up for our daily email digest
Two more wards become Reform vs Lib Dem in council by-elections
Churchill and Lloyd George: Rivals for Greatness - new podcast series
The Joy of Six 1558
Climate change ushers in record-breaking July
"Making Liverpool a global force," Liverpool Lib Dem proposals for regenerating our economy
Cassie
Peter Chambers
Rif Winfield
Nick Baird
Steve Trevethan
13 Comments
I confess that I know less than I should about this whole topic but I sense that it will indeed shape the future and we should be talking about it rather more than we do. So, one or two questions. I assume that the leaders in AI technology are currently the USA and China ? Clearly we don’t want to be reliant on those nations for our AI, so does this mean that we need our own AI companies with our own data centres or can this be achieved in partnership with our European neighbours ? If we need to build these data centres in the UK how will we deal with the inevitable protest groups who will spring up in response ? If the plan is for government to buy significant stakes in the AI industry, what will this cost, and can we realistically afford it ? Navidia has a market value of $5 trillion. AI does not come cheap. Incidentally, these are genuine, not rhetorical questions.
“AI” means “a computer program with a marketing budget”, and in modern terms that means “a largely useless but extremely well-marketed chatbot”. Sure, they *promise* “unprecedented prosperity”, but they have no way to deliver it – all of the benefits are conveniently several years off if they just get a few trillion more dollars to fund it.
Wasting taxpayer money on a bunch of “AI” companies which are not and never will be profitable is a terrible idea. The few big names in the field are essentially being propped up by some extremely dubious financial deals, a massively-inflated investment bubble, and the big tech companies (of which the UK and Europe has no equivalent) essentially donating them huge amounts of server time.
Most importantly, the current round of technology has ridiculous energy and water costs, is causing massive inflation in every other technology device via monopolising RAM production, is built on mass plagiarism and disrespect for copyright and artists, and as you point out is designed and structured for the benefit of a few ultra-billionaires. A UK-owned version of that isn’t an improvement.
The “AI” companies are causing severe problems for the rest of the tech industry. Hardware prices for RAM, storage, PCBs are increasing as the techbros buy up everything for data centres that haven’t been built yet (and probably won’t). Some hardware companies won’t have capacity for consumer tech for at least two or three years. The “AI” tools are generating unchecked code that is causing outages costing companies like Amazon millions in lost revenue. Some smaller software projects have had to close because they’ve been swamped with “AI” generated junk bug reports and support issues. Early adopters like Ford are having to re-recruit senior developers at higher salaries to come in and fix the problems. Best of all, at some point the “AI” companies are going to have to pass on the real costs to customers or go bust – it’s a global game of chicken amongst them as to who will blink first.
I think the more important question is how to protect the economy from when (not if) the “AI” bubble bursts. OpenAI is probably going to be the first to collapse – they’re losing market share fast, only raised enough funding to last another year (rather than the 5 they needed), and now Apple is suing them for alleged theft of their tech. Google is facing lawsuits for breach of licensing contracts and copyright theft from US publishing companies, which could open the floodgates for other such lawsuits against all the “AI” companies.
Good points about alleged copyright theft. Looking at the numbers for “trillion parameter models” it would seem that only pouring all the written material in the world – tagged by low-wage labour or earlier models – would be enough for training. The vendors are very secretive about this.
A wider point from the original post is that many products on the market had the costs of the earlier risky phases of R&D paid for by the tax-payer. A quoted example is the mRNA technology used by Pfizer et al were 97% paid for by the tax-payer in the USA.
Somehow this is transferred for free to the stockholders at the IPO.
A counter-example is the Norwegian Sovereign Wealth Fund. While Britain’s North Sea oil was used for tax cuts and current account spending, Norway invested the money as capital.
@ Peter,
“North Sea oil was used for tax cuts and current account spending, Norway invested the money as capital.”
A couple of points:
1) Norway and the UK had similar levels of oil reserves in 1970, but the population of Norway is less than 10% of the UKs.
2) The purpose of the Norwegian SWF is to export capital rather than invest in Norwegian industry per se. All the big exporters use SWFs to keep the value of their currencies rising too much. This can severely damages the home industrial base and is sometimes known as the Dutch disease.
So we could have had a SWF in the 1980s and the Tories are certainly mainly to blame for what happened. But the rest of the UK population is partially responsible too. We’ve always had a higher-the-better mentality when it comes to the value of the pound. We treat it almost as if it were a “UK Ltd” share price. Other European countries are much more concerned to keep their exporters competitive by keeping down the value of their currencies.
So instead of relying on our industrial base to generate our prosperity Mrs Thatcher and the Tories used oil revenues to fund her class warfare and de-industrialise the North and other regions. If this hadn’t been allowed to happen I’ve no doubt we’d still be in the EU.
We are both saying the same thing effectively but this is a slightly different take on it.
Oh dear. The UK is actually doing quite well for AI firms and investment here – would the state taking over some of the shares make that more or less likely to continue. We need AI firms to prosper, make profits and employ people – then the state will get tax revenue from them.
The difference between AI firm and the Alaskan oil reserves is that unless the oil price plumeted there were bound to be profits to be shared.
Its also far far too early to say that the benefits from AI are going to be concenretrated in a few firms.
@Cim – before you dismiss the effects of AI perhaps you should have a look at the work in science – for example on protein folding for which the founders of Deep Mind won the Nobel prize
@Simon McGrath
That is exactly the marketing trap which AI does, I’m afraid – the pretence that any “computer does something either superficially or genuinely impressive” is therefore all work towards some “just a few years away, honest” unified “computer is intelligent” goal, and is therefore all equally worthwhile.
The technological type of computer program behind the protein folding research – which was a genuinely good piece of work – has only two things in common with the “LLM” and “diffusion” technologies currently being hyped up across the world.
1) It runs on a computer
2) It has a marketing budget which finds it useful (or at least found it useful at the time – a lot of the actually useful stuff is going for “machine learning” to avoid getting tarred by association nowadays) to call it “AI”
And the article seems to be primarily talking about the super-hyped chatbot stuff. The protein folding stuff is really good work, but it doesn’t “promise unprecedented prosperity”, it promises some incremental progress over the next few decades to a better understanding of science and maybe some practical and valued advances in healthcare.
I think that ownership instincts are right, but a fund buying shares in US AI firms makes us shareholders, not builders. It’s a dividend, not an engine. The harder question is where Britain can actually own capability. It is clear now that we won’t out-spend the US on frontier models or match Taiwan on chips, so let’s stop trying. We should back the niches that are genuinely ours and genuinely matter: AI built on our own NHS data, compound semiconductors in South Wales, defence and secure compute, and proper access for small firms.
And we could start by not selling the crown jewels – UK invented Arm and DeepMind, then handed the ownership abroad (Japan and Google).
A short article on the Today programme this week said that in the UK employers were tending to use the GPT-LLM technology to lower costs, for example by sacking “duct taper” roles. By contrast US employers were alleged to be using the technology to swap people around and improve revenues, and hopefully profits.
This sounds plausible as the UK continues with Declinism and the US pursues profit “by any means necessary”. Which includes buying up profitable UK firms using private equity.
We work for them ?
If AI is the sum of all human knowledge, then those private sectors entities that make use of it should pay a usage fee – personally I’d direct the proceeds towards job creation & organisation sponsorship in climate change response
Many US companies are laying off workers to replace them with AI, then finding out that AI doesn’t work as well as humans for most tasks, and are begging their ex-employees to return.
AI is indeed massively over-hyped, as @John C and @cim have pointed out. It is a bubble and it will burst, and once the dust has settled AI (or machine learning) is likely to focus on the useful scientific endeavours rather than the generative and chatbot stuff. It’s like the dot-com bubble — e-commerce didn’t die, it just focused instead on what was useful and realistic.
Jumping on the AI bandwagon may be profitable in the short term for many companies, but this doesn’t mean the results are at all desirable. Publicly listed companies tend to focus on quarterly profits rather than long-term growth. So they replace essential employees with bots to save money in the short term, leaving the resulting problems for the next quarter(s). These include trying to rehire the employees they fired, as I noted above. Newspaper companies are sacking journalists resulting in their publications being full of AI slop. It’s virtually impossible to get one’s social media profile back up after it’s been wrongly suspended by a bot, because there’s meaningful appeal mechanism to get the decision reviewed by a human operator. AI can’t tell truth from fiction, hence hallucinations. Certain things require human judgement, and always will.
AI can’t even judge whether text is AI generated — even deciding something is AI because it uses proper punctuation. My em-dashes are human generated. ———