The following article by Mike Morrissey was first published in Unity, the weekly paper of the Irish Communist Party.
TONY Blair has been on a mission to save the Labour Party, from itself. He has form as a saviour; between 2007 and 2015 he was the Quartet's Middle East Peace envoy – that went well. His new venture takes all the arrogance and unselfcritical confidence he demonstrated at the Iraq enquiry to refashioning the party he led for ten years in which, to be fair, it won three general elections. Now he brings his enormous, self-declared talents to how the UK should be governed – The Labour Party is playing with fire over its future and the future of the country (The Tony Blair Institute for Global Change, 20/05/2026). One key recommendation – an uncritical alliance with Trump in a new age of geopolitical uncertainty – is hardly worth discussing. The word 'Greenland and the phrase 'Board of Peace' say it all. Another involves fully embracing artificial intelligence to make the UK economically successful. Thus, he proposes:
reorganising the whole of government around the harnessing of the 21st-century technological revolution. All governments for the foreseeable future will govern in the age of AI. Those which understand it will see their countries prosper; those which don't, won't.
Actually, the condition of the UK economy may be even worse than he imagines. It doesn't take an economics GSCE to recognise the need for radical policy change. But, is AI the mechanism to deliver that mission?
Some context: proclaiming a 'new thing' as the wave of the future has certainly been his 'thing'. Twenty years ago it was globalisation and financialisation. Then, there was climate-change scepticism. An Institute report (The Climate Change Paradox: why we need to reset action on Climate Change, 2025) suggested that the majority, scientific view was 'riven with irrationality', i.e. he knew better. The same theme continues in his latest effort where he dismisses net-zero targets – unsurprising really given that Saudi Arabia has been a major client of the Institute.
Nor is he a neutral observer of AI. Between 2021 and 2023, his institute took $130 million from Larry Ellison of Oracle (deeply involved in AI infrastructure) and has pledges for a further $218 million with some characterising the Institute as Oracle's 'sales engine' (Jonathan Freedland, The Guardian, 29/05/2026).
Lastly, his intervention was, at least in part, a response to criticism of his own government – Andy Burnham, for example, claimed that Britain has been in the grip of neo-liberalism for 40 years. Blair was clearly piqued that he was included.
Even so, it's worth asking whether AI can deliver the benefits Blair envisages, what will be its likely impact on the economy and jobs and what are the implications of the extreme concentration of its ownership in a small band of 'tech bros'?
AI Capabilities: epoch changing or hype?
While AGI (artificial general intelligence) seems, like fusion power, to be always just a few years in the future, the latest models are clearly impressive. Forget the cute chatbots or the multiple inaccuracies, not to say the dangerous stuff Musk's Grok generates. Coders love Anthropic's Claude Code even if it may eventually make them redundant and it's new model, Mythos, was judged too dangerous for general release until some of the big players, like Microsoft and Google, tested it for dangers to their own systems. It was reported recently that AI was crucial in developing an antigen for a whole family of, rather than just one coronavirus.
Anthropic should also be praised for rejecting the Trump administration's attempt to use Claude for mass surveillance and autonomous weaponry – just imagine Hegseth in charge of swarms of autonomous killing machines.
Most judge that Chinese models are, at most, only months behind the best in the US and there is evidence of more widespread diffusion within that economy. Blair thus seems right in asserting that AI systems will be a systemic feature of contemporary economies.
It's still hard to distinguish between the hype put out by AI companies and the real potential of AI, particularly since very few really know what's going on in Silicon Valley labs. One assessment, however, is scathing (Servaas Storm, Russell's Teapot: dispatches from the final stage of the AI bubble, Institute for New Economic Thinking, 17/04/2026). Even the most advanced models are non-sentient, essentially 'predictive pattern matchers', where the 'hallucination' rate (errors) increases with 'out-of-distribution-data' (other on which they have been trained) – sometimes as high as 50%. This problem is structural and cannot be resolved by scaling. A survey of 6,000 firms that had adopted AI found no measurable improvement in productivity within a three-year period. AI firms are failing to make profits from their LLMs (large language models) and may never do despite huge government contracts – for every dollar of income, Open AI's loses 1.2 dollars; Anthropic only broke even in the second quarter of 2026 (Aisha Down, The Guardian, 03/06/2026). Thus, the rush to invest in data centres may be an unsustainable bubble.
In short, while impressive things have been done using AI, it may be more limited than its evangelists claim.
FOBO
There remains, however, widespread concern about the technology's disruptive, if not destructive, impact on employment, summoned up be the phrase 'fear of becoming obsolete' (FOBO). Predictions about job creation and job loss via AI vary significantly. The Economist published a leader article on May 14th entitled 'Prepare for the AI jobs apocalypse'. Goldman Sachs estimated that between 4 and 22 million jobs in the US face 'displacement risk' (www.axisintelligence.com/ai-job-displacement-analysis-2030-prediction).
The World Economic Forum (30/04/2025) claimed 'the rise of AI-powered tools threatens to automate as many roles as it creates' and suggested it would affect 50 million jobs in the US. Its 2025 Future of Jobs report found that almost half of employers expected to reduce their workforce where AI could automate tasks. The Institute for Public Policy Research is more cautious – 'Instead of ushering in a "jobs apocalypse," AI technologies have the potential to raise productivity growth and improve people's experience of their jobs' (Strike while the AI is Hot, May 2026) – that, however, depends on the implementation of policies to both ameliorate risk and augment worker capacity to influence the changes that affect them.
A Few ?Good? Men
Keynes once predicted that productivity increases by 2030 would allow a universal 15 hour working week. That is more than possible today but not within AI's current ownership structure. There are already worries that the investments needed to create larger and larger data centres are being partially funded by cutting jobs altogether.
Up to now, most companies have been private with investment self-funded. No longer; some of the biggest are going public this year (possibly adding $4 trillion to US listed companies capital value), with IPOs designed to retain control. This model, called 'funder control' (it negates the one-share-one-vote principle) pioneered by Meta and Alphabet, ensures 'Founders retain voting control despite owning a minority of the equity, in effect silencing public investors' (Gill Whitehead, The Economist, 21/05/2026).
It is also a core feature of the SpaceX offer to raise $75 billion (capital value $1.75 trillion, 92 times its annual revenue) on future plans for space tourism, energy production/manufacturing on the Moon/Mars and asteroid mining (Dan Milmo, The Guardian, 21/05/2026). It proposes to build data centres in space (orbital compute) which would avoid both regulation and taxation. In short, the world's potentially most disruptive technology is under the control of a small band of 'tech bros', at least one of whom appears increasingly unstable – what could go wrong?
Paradoxically, this tight concentration of ownership and the possible disruptive impact on jobs strengthens the case for collective ownership and socially planned change. Already there have been calls for Universal Basic Income if unemployment rapidly increases. The challenge is to develop a collective response to ensure this technology is developed for the common good – something missing in the Blair piece.
Yesterday's Men (and one woman)
Aditya Chakrabortty (The Guardian, 04/06/2026) reports the public think Tony Blair less worth listening to than Boris Johnson. Yet, in the week his essay was published, the national newspapers gave it three times more coverage than that of the Home Secretary and Foreign Secretary combined. Bereft of a strategic vision, the UK Government seems to believe that salvation can only be found by resurrecting New Labour figures. Brown and Harman have been brought in to advise. Others, like Johnathan Powell, are already there. Mandelson even sent copious messages, before his downfall, revealing his contempt for most of those in government. Blair did win elections. He also saw the loss of 'a million manufacturing jobs, a huge fall in council housing and a historic financial bubble' from which the British economy has yet to recover. Truss's disgraced ex-chancellor, Kwasi Kwarteng, asserted on GMTV that he agreed with Blair's essay – perhaps that's Blair's real audience.