In January, ahead of their attack on Iran, the US Department of War announced a new Artificial Intelligence Acceleration Strategy to usher in an “era of American military AI dominance”. Secretary of War, Pete Hegseth, promises to “eliminate legacy bureaucratic barriers” in a plan to guarantee warfighting advantage, and “eradicate woke DEI from AI capabilities”. The full plan reads like a fascist’s wet dream, but it’s real, it’s happening fast, and it’s in the hands of people who in the past few days, with no regard to morality or widely understood ‘rules of war’, have bombed schools, and torpedoed boats in international waters, leaving their crew to drown.
The AI strategy is framed as essential because we are in a race. It’s a characterisation very reminiscent of the nuclear arms race – a technology so destructive that international treaties, backed by rigorous regulation and independent inspection, have sought to control ever since the United States used it to kill hundreds of thousands of citizens (including tens of thousands of children) in Japan.
According to a recent report from the independent Ava Lovelace Institute (named after the 19th century female maths pioneer), the government is way out of step with public concerns over the rapid unregulated expansion of AI into many aspects of infrastructure and society, including the military.
A scientific poll of a randomly-selected but sortitioned cross-section of the public (i.e. demographically representative) assessed the UK public’s awareness, experience and views on the benefits and concerns associated with AI technologies. It found a massive majority (around 9 in 10 people) wanting the development of AI to prioritise fairness, safety and benefit to society, over economic gains, speed of innovation or international competition.
In contrast, Labour has made slow progress on its election promise of statutory regulation, and instead announced policies supporting rapid expansion of this sector, with its associated energy needs.
So last weekend, a new people’s movement, Pull The Plug, joined forces with Pause AI in organising a march in London, calling for similar regulation of Artificial Intelligence (AI). Billed as the March Against The Machines, they are not against AI itself, but against its unregulated use by governments and corporations which are more likely to harm than benefit the public.
One of the issues covered is how young people and racialised communities are particularly adversely affected by social media, algorithms and AI. Mad Youth Organise are calling for a tax on big tech in order to fund mental health programmes. Also supporting the march were Black Sox, an organisation working for global equity for Africa and the Global South. Around a couple of hundred people visited the HQs of OpenAI, Meta, Deep Mind and Google, hearing speeches at each site.
Dr Rachael Kent is a Senior Lecturer on digital economy at Kings College London, who successfully took Apple to court in the UK, winning a historic £1.5 billion for 36 million App store consumers. She spoke about the structural dynamics that shape pricing, which also shapes our knowledge environments every day. She gives as an example, that without robust regulation, health misinformation becomes a really important commercial advantage for these businesses, and we’ve seen the spread of misleading anti-vax and bio-hacking information to millions of users, simply because algorithmic amplification actually incentivises harm because it circulates and attracts attention.
Dr Kent argues that government must treat digital platforms as information infrastructure and must resource regulators to properly audit algorithms, demand transparency, and impose meaningful sanctions and penalties when harms occur. She warns that when information power is concentrated, accountability weakens, and harm spreads. We must refuse to accept that the current concentration of control is simply the price of information.
Dr Tom Stoneham, Professor of Philosophy at York University, talked about the power consumption of the new breed of AI data centres. OFGEM recently announced it had received a request for an additional 50 Gigawatts of energy to power the next generation of 140 new data centres in the UK. To put some perspective on that, Dr Stoneham revealed that on the coldest day this February, the UK hit a peak of 45 Gigawatts consumption. The government is pushing ahead with plans for more than a dozen privately-run small nuclear reactors across the UK, with many in the north-east. Hartlepool’s aging EDF-run reactor was put under special measures by nuclear regulators last year after a series of site incidents, but the nuclear industry is pushing ahead with massive expansion plans. Some are also warning that the grid system itself will come under strain, and that the large number of prioritised connection requests will crowd out renewable energy schemes already waiting for grid access, thus harming the UK’s attempts to meet emission targets.
CEO of Pause AI Global, Maxime Fournes, worked in the AI industry for twelve years, but realising the dangers, he did “the only thing as a decent human being” and stopped. Although he has no illusions that AI companies will not respond to lobbying, he is encouraging whistleblowers, and is working hard to educate the public that working in AI is an anti-social activity.
There is no doubt that a powerful new anti-democratic force is emerging – a mixture of money, politics and information control in the hands of super-rich tech oligarchs, but the odd thing is that even they have misgivings. In a recent interview Elon Musk said “If I could, I would certainly slow down AI and robotics, but I can’t. It seems to be advancing at a very rapid pace, whether I like it or not.”
Likewise, former comms chief at Google DeepMind, Dex Hunter-Torricke, has called on governments to act, warning in his essay ‘Another Future Is Possible’ that “the path we are currently on leads to disaster.” His main thrust is economic – he sees job losses and speedy changes in economic structure as inevitably leading to a world in which an elite class live long lives of healthy luxury, while the global majority suffer a permanently curtailed existence in poverty with little or no healthcare access, nor any political power.
Two safety researchers at OpenAI (ChatGPT) and Anthropic (Claude) quit their jobs last month, making public a whole list of concerns. OpenAI has recently disbanded its “mission alignment team”, tasked with ensuring the company’s direction benefited humanity, and it has also signed a new deal with the Pentagon after Anthropic raised concerns over military use of its AI models during Iran strikes, and ended up in a spat with President Trump.
Google’s old motto “don’t be evil” has gone by the wayside in many different ways over the years, but most recently last month they simply deleted a clause in its policy that prevented DeepMind from use in surveillance, weaponry and other technology intended to hurt people.
At the same time, CEO Demis Hassabis has actually announced that he supports ‘a pause’ in principle. This brings us back to that comment by Musk – despite misgivings, what people might want, and what will actually happen, are two different things, and unless real pressure comes from somewhere, market forces and technological advances will usher in a dystopian regulation-free future rather than a potential utopian one.
A coalition of groups are calling for Citizen’s Assemblies to decide how AI will be used and who will control it. The protest in London ended in a church hall in Bloomsbury with its own model people’s assembly, organised by Assemble, where dozens of people interacted and developed a strategy for action and movement-building that Pull The Plug will promote in the coming months.
AI technology is expanding fast, and many of its potential consequences are as yet unknown, so it is one of the most important issues that ordinary people need to have a voice in. Pushing for new ways to empower that voice may just help solve other major issues facing us all.
