Frank Hansen discusses a sober account of the dangers and possibilities of AI
Confused by the hype about whether AI will destroy humanity or solve all our problems? Karen Hao’s brilliant book (Empire of AI – Inside the reckless race for total domination) provides a clear view of what’s really going on inside Big Tech as the AI companies compete in a race to market, based on the drive to make super profits.
Of course, AI can be highly beneficial. Recently, neurosurgeons performed the first successful AI-assisted operation to remove a brain tumour in partnership with London University. However, is the massive wave of AI investment being driven by Silicon Valley likely to be beneficial for humanity? Hao’s book focuses on the rise of Sam Altman and OpenAI. It exposes the lies we are being fed and uncovers the damage that Big Tech is inflicting on society and the environment in the real world today.
There are many pathways for developing AI, and alternative, democratic models are entirely possible. Yet only one (Generative AI), has been prioritised by Silicon Valley as the most profitable. Unlike traditional AI, which analyses data or makes predictions, GenAI produces new outputs by learning patterns from vast datasets, which can then be monetised. To develop these models, huge resources are required – finance, labour, information, energy, water, computers, hyperscale data centres. This is known as the “Gospel of Scale” – Altman’s belief that building AI requires exponential increases in data and computing power.
OpenAI was founded as a safety-first, non-profit research lab. To fund the relentless drive for “scale”, it transitioned into a profit-driven entity, accepting billions from Microsoft. This transformed the organisation: intellectual property was locked away, and safety priorities were sidelined in a breakneck race to market. All leading AI firms now follow the same model. Hao argues they act like modern-day, neo-colonial empires aiming for global dominance by:
Resource Seizure: Using global data, intellectual property, and natural resources sometimes without consent or compensation. This is obvious in the creative industries where internet data is “scraped” without permission or offering compensation. Creators’ work is extracted to train algorithms designed to automate their professions, resulting in job losses and poorer terms and conditions. Huge data centres, needed for training large models and powering global cloud networks, are now appearing. These consume large amounts of energy and water. They threaten local supplies and contribute to global warming. Local protests are now escalating in the US, Chile and the UK.
Labour Exploitation: Using low-wage workers in the Global South to sort vast amounts of data, some of it toxic, to build highly profitable products. Chatbots do not naturally understand human speech or distinguish safe content; they require human feedback. Hao highlights the exploitation of “digital sweatshop labour” in Kenya and elsewhere.
Knowledge Monopolisation: Funnelling top researchers and technical infrastructure into private, opaque corporate vaults. Monopolising science and holding back more benign alternatives, and restricting independent scrutiny
Ideological Justification: Using a “grand mission” to excuse immediate harms. Lobbying governments to ensure that regulatory frameworks are “acceptable”. Hao argues that Altman uses AI as a promise of salvation, suggesting that future AI will cure all human diseases, etc. Simultaneously, executives weaponise “AI safety doomerism”, warning that AI could become an existential threat to humanity. This shifts the focus away from immediate harms towards regulating speculative scenarios. Big Tech has argued for regulation, but only on their own terms. Altman has proposed an international body – led by the US, bypassing the UN – to regulate AI. The Tech Titans have ulterior motives; we cannot trust what they say – regulation needs to be independent and under democratic control.
US AI policy has been captured by the claim that “over-regulating” American firms would hand a victory to China. In the UK, tech lobbying has also shifted policy towards “pro-growth”, emphasising the danger of “falling behind”. Plans for large data centres are now being implemented across the country. The Government’s promise of AI Safety legislation has been dropped. A Lords amendment to the 2025 Data Act requiring firms to disclose the copyright data used was defeated in the Commons. Recently, a data centre development in Brick Lane was approved on appeal by Angela Rayner, overriding strong local objections. In contrast, the EU has a strategy of robust regulatory enforcement, combined with a sovereign industrial policy.
Protests over data centres and concerns about “runaway AI” provide an opportunity to press for change in Government policy, towards public regulation and accountability, to ensure AI is developed safely in the interest of humanity. The belief that Big Tech’s AI vision will automatically deliver progress and “growth” needs to be scrutinised and tempered by the possibility of other threats – the last tech “dot.com boom” in 2000 fuelled uncontrolled speculation and was a contributory factor in the 2008 financial crash.