AI’s rapid insertion into the economy and society at large is deepening the crisis of an already fragile political system. Grave dangers include the further undermining of human rights, AI platforms taking over state functions and mass unemployment. At the same time, opportunities for a progressive, shared future present themselves if we take up the challenge of ownership and control.
The leap in technology to artificial intelligence is undoubtedly an historic step forward in social history. Uniquely, the accumulated knowledge of humanity is made available in the form of a digital tool and is already playing a critical role in the productive process.
For sure, this is being done in the name or efficiency and profit at the expense of workers and society at large. However, instead of standing back in awe or simply trying to block AI, if we focus on how to transform its ownership and control, the potential for common benefit is unprecedented.
The technology is developing at a phenomenal pace. In recent months, it has moved on from large language models (LLMs) that respond to user prompts to Google and other providers announcing that the era of Agentic AI has arrived.
Agentic AI models are given actual tasks by humans which they then complete autonomously. It is a system that, according to an industry analyst:
- is provided with a goal or task (not just a single prompt)
- plans the steps required to complete it
- takes actions – calling other applications, browsing the web, writing and running code, sending messages
- observes the results of those actions
- adjusts and continues until the task is done.
Shock waves are already running through the economic system before Agentic AI rolls out. Banks are laying off thousands of staff, led by Barclays who announced that 5,000 jobs would go largely because of AI. Virgin Media02 revealed a similar level of redundancies. Cloudflare, a cybersecurity and internet infrastructure provider announced that 20% of its global workforce would lose their jobs.
Estimates of job losses vary, but here’s the range:
- Half of entry-level white-collar jobs are expected to be lost by 2030
- The UK is predicted to experience twice the international average of job losses due to AI, with a net loss of 8% over the past 12 months
- AI-driven job disruption is expected to account for 22% of jobs globally by 2030
- 75 million jobs are projected to be replaced globally, with up to 270 million potentially replaced entirely by 2030
- 300 million jobs are expected to be exposed to automation, with 92 to 270 million of those potentially replaced entirely by 2030.
In a worst case scenario you could have about 8 million jobs [in the UK] being negatively affected by this [AI].
Carsten Jung, of the IPPR think tank,
If job losses are as anywhere as catastrophic as the forecasts, then the state will not be able to cope financially. You can’t imagine even modest benefit payments being sustained by governments when a large proportion of the working population is jobless. The loss of state revenue from unemployment and the corresponding reduction in consumer spending and tax revenue will undermine already shaky government finances.
With the fabled mantra of ‘growth’ failing to materialise, borrowing has filled the revenue gaps. Nearly 80% of borrowing by governments in the major economies will go to refinance existing debt, which is becoming more expensive as interest rates rise. It’s like using a global credit card to raise cash to pay your debts – it’s simply not sustainable.
With financial markets showing signs of distress, borrowing by AI corporations has squeezed out other firms from the market. AI-related debt accounts for 25% of all corporate bond issuance, up from 4% in just two years. Borrowing rose from near zero in 2024 to $500 billion within the first nine months of this year. The Bank of England has warned that a failure in one or more large AI borrowers could tip the entire financial market over the edge. The run on Google and Intel shares in the past week following the defeat of the Republicans in the US mid-term elections is perhaps an indication that a crash is already under way.
Meanwhile, the penetration of state functions by the tech corporations is well under way, posing a distinct threat of its own. AI-driven platforms operating deep inside the NHS, military and policing are a reality. The Burnham government has decided to stick with Palantir’s health service contract, which is due for renewal next spring, despite the corporation’s close involvement in the Israeli genocidal attacks in Gaza and the brutal rounding up of migrants in the US.
Palantir’s founder, Peter Thiel, has questioned the usefulness of democratic political systems, claiming that tech companies can “unilaterally change the world” and called AI an “incredible alternative to politics”. Couldn’t be clearer than that. Under Thiel’s nightmare totalitarian vision, shareholders would replace voters. RIP representative democracy, already in meltdown around the world.
A few senior political figures are conscious of the dangers – without offering any solutions. London mayor Sadiq Khan warned that AI could become “a weapon of mass destruction of jobs” if not properly controlled. Urgent action was needed to “seize the potential of AI and use it as a superpower for positive transformation and creation” and to prevent “a new era of mass unemployment”.
London was “at the sharpest edge of change” because of the “colossal” impact on jobs in key sectors including finance, professional services and the creative industries, Kahn added. Khan has spoken out against Palantir’s involvement in Metropolitan Police activities, but was compelled to withdraw after he was hit by legal action.
There is no sign of the Labour government heeding his warning. The recent pathetic budget, which consigned most of the population to further years of high prices and failing services, did, however, declare significant support for corporate-led AI. Chancellor John Healey revealed £700m worth of plans to prioritise data centres for electrical grid connections, which will override local opposition to these water and power-hungry infrastructures. Offering up a huge dose of nationalism, he said AI would be “built on sovereignty and security” so that British people can benefit from the ”huge potential upsides of AI”.
When the state is unable or unwilling to deliver reforms, is economically and financially in difficulties and turns to authoritarian rule, frustration grows and protest can take violent, unpredictable turns – expressing anger without a clear view of how to proceed. Flashpoints in France, Germany, Italy and India, led by new generations, indicate the collapse of state authority.
In the United States, a record turnout in the mid-terms produced a massive vote against corruption and autocratic rule and robbed Trump of his control of Congress. In other words, there are vast social movements of people springing up everywhere, taking different forms ranging from voting to street battles with state forces.
AI is a phenomenal development which could help create a society where personal, cultural, physical and intellectual development is the norm, rather than what job can I get, what happens when I get fired, how can I afford to eat well, how can I afford decent housing and how will l manage in older age?
However, at present the purpose of AI is to accumulate profit, to pay off borrowing and reward shareholders. There are a variety of approaches aimed at limited the power of the AI corporations. They include state regulation (whatever that means), and more effective ‘guard rails’ built by AI firms themselves to prevent rogue operations threatening humanity. .
Others advocate challenging the Googles, Anthropic and Microsoft of this world with publicly-owned, rival AI firms. Not-for-profit cooperative enterprises are discussing alternatives to big tech. Direct action campaigners advocate Pull the Plug or Cut the Power as a way forward in opposition to the massive expansion of data centres and the influence of AI corporations. The central issue of ownership and control remains in the background, however.
To realise AI’s potential requires liberating it from its corporate form of ownership. To achieve that requires a mechanism for converting Anthropic and the others into co-operative, not for profit types of enterprises. That, in turn, takes us in the direction of political, state forms that can carry this through.
New forms of democratic power are being explored around the world. In Serbia, a decentralized student-led movement defeated the far right, nationalist government, only to find that the state remains in the hands of the old guard. They have asked them to resign their posts. The movement is confronted with the task of remaking the Serbian state from top to bottom.
The goal everywhere should be the creation of new, more advanced forms of democracy, moving ahead from representative systems that are dying in front of our eyes, opening the door for dictatorial rule overseen by AI corporations like Palentir.
For AI to work for the people, it has to be owned and controlled by them, exercising power directly through bodies like assemblies. Then we could unshackle its tremendous potential and begin to truly revolutionise our futures.
This post was inspired by a recent symposium at Cambridge University on AI and human intelligence, organised by the International Friends of Ilyenkov.
AI won’t/can’t do anything unless it’s prompted, told to by humans, which effectively means those who control the systems. Their use by the Israeli and US military is another strong argument for taking it out of Big Tech’s control – and for remaking the states that authorise what is essentially digital murder.