Guest Author: Raphael Dennett – University of Exeter
Every business leader in Britain faces key choices in the next twelve months, and they are perhaps not the ones most of them are prioritising. What if the key risk is not falling behind on AI, it is what happens to your business, your people and the country if businesses get the response to AI wrong? According to the British Chambers of Commerce’s latest research, 54% of British firms are now using AI. 95% of them say it has had no impact on their headcount. Both of those statements are true, and if you find them reassuring, you have not read the rest of the data. In a recent interview in February 2026, Mustafa Suleyman, the head of Microsoft’s AI division, told the Financial Times that most white-collar professional tasks will be fully automated within twelve to eighteen months: lawyers, accountants, project managers, marketing teams. His exact words were that AI is approaching human-level performance on most, if not all, professional tasks. This is not a researcher doing some armchair speculation, he is one of the people building the products designed to make it happen. Some may disagree with his timeline, but almost nobody disagrees with the direction of travel, and it is one that every British business leader ought to be taking seriously right now.
He is not alone. Sam Altman, CEO of OpenAI, has argued that AI agents will effectively join the workforce as autonomous contributors to economic output within months, we are already starting to see this happen. Geoffrey Hinton, the Nobel Prize-winning computer scientist who helped build the AI technology now reshaping the labour market, warned in September 2025 that AI will create massive unemployment and a huge rise in profits, and that the gains will flow to a small number of people rather than to the majority of society. Researchers at the investment firm Citrini recently published a scenario that the financial markets found uncomfortably difficult to dismiss: AI job displacement outpacing the ability of companies to adapt, a collapse in consumer spending, and government tax receipts disappearing at a catastrophic rate as productivity gains flow to capital owners rather than to households. US tech stocks fell sharply in the days that followed. The investment experts whose job it is to price future risk sold sharply on the back of a simple thought experiment, we should probably pay attention to this reaction.
Most of the data coming through in early 2026 suggests that British business has not been paying nearly enough attention. Gardiner and Theobald found that 97% of British organisations report at least one significant AI skills gap, with a third saying those gaps are already hurting their ability to meet business goals. The CBI’s AI Skills report, published in January, found that many businesses are experimenting with AI without the training, the guidance, or the capability to scale what they are learning. A Parliamentary briefing in February found that just 21% of UK adults can explain AI in any meaningful detail, and only one in five people in work feel confident using it. Rather than being just projections or abstract guesses, this is the state of the British workforce today.
The BCC’s new report, Powering Productivity: AI and the Future of UK Work, published in March 2026 in partnership with Atos, fills in the picture in considerable detail. That 54% adoption rate is up from 35% in 2025 and 23% in 2023, a pace of change that would have seemed implausible even two years ago. However, around one in ten firms have moved beyond generic tools like ChatGPT and Copilot into deeper, bespoke AI integration, and it is in this group that the economic consequences are already showing: approximately one fifth report staffing reductions attributable to AI, and bespoke AI adopters are roughly three times more likely to have restructured job roles. The firms furthest along the adoption curve are the ones already cutting and reshaping their workforces, and that should tell us a great deal about what the next twelve months will look like for everyone a step or two behind them.
There is something else happening that the headline employment data does not capture yet. Headcount may not be shrinking across most British businesses yet, but hiring has effectively frozen. UK job vacancies dropped below 700,000 in January 2026, the lowest level since the pandemic. Graduate roles are down 45% year on year, and each vacancy now attracts hundreds of applications. The Randstad Workmonitor 2026 report found that 38% of employers plan to hire fewer graduates specifically because of AI, and the chief executive of Robert Walters described this in March as the longest hiring downturn his industry has ever experienced, longer than the financial crisis and longer than the pandemic. The CIPD’s latest Labour Market Outlook confirms that one in six UK employers now expect AI to shrink their workforce over the next twelve months, with the risk highest at larger private sector firms. The labour market is contracting through the slow closing of opportunities, fewer new hires, fewer replacements, fewer entry points, and that pattern almost always precedes outright headcount reduction and this is only going to get more pronounced once AI adoption starts embedding.
Assuming that this is an AI skills issue is one of the mistakes that most corporate training responses to AI are currently making. Morgan Stanley’s research, published in late January, found that UK firms using AI for at least a year report net job losses of 8% over the past twelve months, the highest rate among major economies surveyed and roughly double the international average. American firms reported similar productivity gains from AI adoption but created more jobs than they cut. The difference doesn’t seem to be the technology itself but rather it is the strategic decisions that businesses are choosing to make as they start the adoption process. American firms used AI to grow and develop new innovations and capabilities while British firms used it to shrink, and the consequence of that choice is that we are not expanding our economy with this technology but contracting it. As the Boston Consulting Group argued in April, the nearer and broader effect of AI is not simple elimination but redesign. Its modelling suggests that over the next two to three years 50% to 55% of jobs in the United States will be reshaped by AI, while 10% to 15% could be eliminated over roughly five years. The immediate challenge is not only redundancy but the restructuring of work, work structures and work strategies. Rachel Fletcher, head of EMEA sustainability research at Morgan Stanley, described the findings to Bloomberg as an ‘early warning sign’ of how AI is disrupting the labour market, and noted that the technology’s impact on employment had ‘come up in a lot of our recent investor conversations.’ It is unusual for investors to consider workforce strategy as a key threat to profitability and business leaders should be considering what this change means for them. What makes this worse is something that is being largely overlooked in the conversation about AI and jobs, which is the question of where the value is ending up, especially in the UK. The AI tools that British businesses are using to replace British workers were not built in this country, the profits from them do not remain in this country, and the new jobs that the technology companies keep promising will be created by AI are overwhelmingly not appearing in this country. Hinton warned in November 2025 that the big technology companies are betting on AI causing massive job replacement because that is where the big money is going to be, and what he is describing, whether he intended to or not, is a transfer of economic value out of the British workforce and into the balance sheets of overseas technology firms. British businesses that respond to AI by freezing or cutting headcount and buying in more software are not just hollowing out their own organisations, they are actively participating in that transfer outside of Britain. If the pricing of compute tokens increases, many of these businesses will be in trouble. The alternative, and it is one that requires a fundamentally different approach to workforce development, is to invest in building the capabilities of British workers to participate in a future economy that will work for us all.
So what are businesses actually doing to prepare their people? The honest answer, for most of them, is not enough and not well. LinkedIn’s 2024 Workplace Learning Report found that fewer than 5% of large-scale corporate reskilling programmes advance far enough to even measure their success, let alone prove that they achieved its objectives. The Harvard Business Review, cited in the same report, concluded that among companies that have embraced the reskilling challenge, only a handful have done so effectively, and even their efforts have been subscale and of limited impact. The Nash Squared Digital Leadership Report found that 59% of UK organisations are not upskilling employees in generative AI at all. IBM’s research found that while two thirds of UK firms report AI-driven productivity gains, only 45% offer any workforce training to support it. What you actually find inside most organisations, if you look honestly, is a split: a small group of self-motivated people driving their own AI learning, and a much larger group who have not engaged meaningfully and are not being reached by whatever programme sits on the company intranet. The motivated, capable few pull ahead while everyone else is either being carried by the organisation’s existing momentum, or trying to protect their position, or slowly beginning to realise that their skills are losing value, and responding to that realisation not with renewed effort but with anxiety and withdrawal.
This is because corporate reskilling is overwhelmingly designed to add new technical competencies to an existing workforce, to teach people new tools and new platforms and new processes. The World Economic Forum put it neatly this month: the bottleneck is rarely the technology itself, it is how work inside organisations is designed. The deeper and more challenging questions, how people approach their work, how they recognise opportunity and mobilise others around it, how they create genuine new value rather than simply executing more efficiently within existing frameworks, are almost never addressed by in-house training. Daniela Amodei, co-founder of Anthropic, was arguing in February 2026 that the very things that make us human, our capacity for communication, innovation, emotional intelligence and collaboration, will become more important and not less as AI capability grows. She is right. But those are not capabilities you can easily deliver in a module or measure in an online assessment. They are professional orientations that develop through practice, over time, in conditions of genuine uncertainty, and they require a learning environment that most corporate training functions simply cannot provide. Skills England recently observed that 71% of the workforce that will be present in 2035 is already in employment. The people who need reorienting are mostly not sitting in lecture halls, they are sitting at desks in offices across Britain, right now, and most of them are not getting the help they need. If the UK really does need to reskill something like 20 million people in just a few years, a few click-through online AI courses will not do it, this is a generational education challenge.
There is a dimension to all of this that is not being discussed with the seriousness it deserves. A badly handled AI transition will not simply displace workers from one kind of role into another. It risks pushing a significant number of working-age people out of the workforce altogether. Britain already has a severe problem here that predates AI entirely: the government’s Keep Britain Working review found that 2.8 million working-age people are now economically inactive due to ill health, an increase of around 800,000 since before the pandemic, at an annual cost of two hundred and twelve billion pounds. Mental health is the leading cause of absence. Mental Health UK’s Burnout Report 2026 found that one in five workers took time off due to stress, rising to two in five among those aged 18 to 24. Now add the psychological weight of AI-driven uncertainty to a workforce that is already under this level of strain. EY’s research found that 65% of employees are anxious about AI replacing their job. University of Florida researchers have gone so far as to propose a clinical framework for what they call AI Replacement Dysfunction, describing the chronic stress and sense of professional obsolescence that is emerging among workers who cannot see a clear path to adaptation. If we do not get the retraining and reorientation of the British workforce right, the consequence will not be a neat transition from old roles to new ones. It will be a further surge in long-term economic inactivity that this country simply cannot afford.
As we have seen, the kind of reorientation that is required cannot be delivered entirely in-house, and this is a conclusion that most corporate training functions find deeply uncomfortable. It requires partnership with institutions that can offer intellectual depth, sustained engagement with uncertainty, and real pedagogical expertise. British universities are an obvious and underused resource. The most effective models emerging from the evidence are not just universities delivering courses to employees but businesses and universities co-designing programmes around real operational challenges, with cohorts drawn from actual teams working on live problems. There is an opportunity here, at the very moment when British businesses desperately need what universities can provide, British universities are themselves in serious financial difficulty, and a properly funded national commitment to workforce reorientation through higher education would address both problems at once. The BCC report’s recommendation that government should lead national efforts to integrate AI literacy across education and lifelong learning points in the right direction, as do its calls for Growth and Skills Levy funds to be directed toward AI training and for the establishment of an AI Labour Market Observatory. But the framing of this challenge is also shifting at international level. In April, the IMF argued that AI should be treated as a macro-critical transition rather than a standard technology shock, and its latest World Economic Outlook argued that labour-market institutions need to help workers reallocate, stay skill-ready, and move across jobs with access to portable benefits and support during transition. That is a much larger and more serious policy task than a few training subsidies and some gentle encouragement to adopt AI. Government also needs to prepare, with considerably more urgency than it is currently showing, for employment shifts that will not be reversed by training subsidies alone, for a lifelong learning infrastructure of a much larger scale than is currently being planned for and for mechanisms to support both businesses and individual workers through a transition that, if the BCC’s adoption trajectory continues, will starting having a serious impact much faster than most policy frameworks are currently designed to accommodate.
What the economy needs is not a workforce that has been trained to use AI tools with greater efficiency. It needs a workforce oriented toward what can be called Regenerative Innovation: the capacity to look at any situation and see not just what is there but what could be there, to recognise where genuine new value might be created, and then to deliver on those opportunities by bringing people with them and making change happen. This is something AI cannot easily do and it is what every British business needs more of, in every role, at every level. If you want a rough formula, the training equation is something like 20% applied AI skills, 30% innovation capability, 50% change management and all of it needs to be framed in terms of human centric culture change. AI does matter, but it is all the capabilities around it that will determine whether British businesses thrive, merely survive or collapse. Even OpenAI, in a policy paper published in April 2026, acknowledged that AI may disrupt jobs and reshape entire industries at a speed and scale unlike previous technological shifts, that gains could become concentrated within a small number of firms, and that workers may become more productive without feeling that they are sharing in the benefits. It even floated time-bound 32-hour or four-day week pilots with no loss of pay, which is a remarkable sign of how far this debate has already moved. This is no longer only a conversation about software adoption. It is a conversation about distribution, bargaining power and the future shape of working life. The businesses that invest seriously in building those capabilities in their people, in genuine partnership with universities and training providers who know how to develop them, will be the ones that find new markets, innovate, create new value, and build the kind of organisations that attract and retain the people who will be properly equipped to continue creating value in the decade ahead. The BCC’s data tells us that the window in which this transition can be shaped rather than suffered is still open. It will not stay open for long. The question facing every British business leader right now is not whether this is coming. It is whether you intend to shape it, or to be shaped by it.
Sources:
- Amodei, D. [Daniela]. (2026, February 7). Interview on human skills and AI [Television broadcast]. ABC News. Reported in: Moran, L. (2026, February 7). Anthropic co-founder says studying the humanities will be “more important than ever.” Fortune. https://fortune.com/2026/02/07/anthropic-cofounder-daniela-amodei-humanities-majors-soft-skills-hiring-ai-stem/
- Altman, S. (2024, September 23). The intelligence age. https://ia.samaltman.com/
- British Chambers of Commerce. (2026, March 18). Half of SMEs Using AI, With Limited Headcount Impact So Far. https://www.britishchambers.org.uk/news/2026/03/half-of-smes-using-ai-with-limited-headcount-impact-so-far/
- British Chambers of Commerce. (2026, March). Powering Productivity: AI and the Future of UK Work. In partnership with Atos. https://www.britishchambers.org.uk/wp-content/uploads/2026/03/Powering-Productivity-AI-and-the-Future-of-UK-Work_FINAL.pdf
- CIPD. (2025, November). Labour Market Outlook: AI and workforce planning. Reported in: Digit.fyi. (2025, November 10). One in six UK employers plan AI layoffs in 2026.
- Citrini Research & Shah, A. (2026, February 22). The 2028 global intelligence crisis. https://www.citriniresearch.com/p/2028gic
- Confederation of British Industry. (2026, January 5). AI skills report: Building an AI-ready workforce. CBI. https://www.cbi.org.uk/articles/ai-skills-report/
- Department for Work and Pensions & Department for Business and Trade. (2025, November). Keep Britain Working Review: Final Report. HM Government.
- EY US. (2025). AI Anxiety in Business Survey. Ernst & Young LLP.
- Gardiner & Theobald. (2026, February). Growing AI skills gap across UK workforce. https://www.gardiner.com/stories/2026/growing-ai-skills-gap-across-uk-workforce
- Hinton, G. (2025, September 6). Interview on AI, unemployment and the capitalist system. Fortune.
- Hinton, G. (2025, November 14). Interview on AI and job replacement. Bloomberg.
- IBM. (2025, October). Race for ROI: UK AI productivity and training findings.
- Institute for Student Employers. (2025). Graduate recruitment trends data.
- LinkedIn. (2024). 2024 workplace learning report. LinkedIn Learning.
- McNamara, S. & Thornton, J. (2025). Artificial Intelligence Replacement Dysfunction (AIRD). Cureus. DOI: 10.7759/cureus.93026
- Mental Health UK. (2026, January). Burnout Report 2026.
- Metaintro. (2026, March). UK Job Vacancies Fall to Lowest Level Since the Pandemic.
- Morgan Stanley. (2026, January 26). AI job cuts and UK labour market. Reported in: Bloomberg.
- Nash Squared & Harvey Nash. (2025). Digital Leadership Report.
- Parliamentary Office of Science and Technology. (2026, February 25). Artificial intelligence and employment (POST Note No. 757).
- Randstad. (2026, January). Workmonitor 2026: AI and graduate hiring.
- Robert Walters. (2026, March). CEO commentary on UK hiring downturn.
- Skills England. (2025, October). A new report charts the path to an AI-ready UK workforce.
- Suleyman, M. (2026, February 12). Interview with the Financial Times. Reported in: Fortune.
- World Economic Forum. (2025). The future of jobs report 2025.
- Boston Consulting Group. (2026, April 3). AI will reshape more jobs than it replaces. BCG. https://www.bcg.com/publications/2026/ai-will-reshape-more-jobs-than-it-replaces
- International Monetary Fund. (2026, April 3). Global economic and financial implications of artificial intelligence: Lessons from a scenario-planning exercise (IMF Note No. 2026/002). IMF. https://www.imf.org/en/publications/imf-notes/issues/2026/04/03/global-economic-and-financial-implications-of-artificial-intelligence-lessons-from-a-574924
- International Monetary Fund. (2026, April 14). World economic outlook, April 2026: Chapter 1. Global prospects and policies. IMF. https://www.imf.org/-/media/files/publications/weo/2026/april/english/ch1.pdf
- OpenAI. (2026, April). Industrial policy for the intelligence age: Ideas to keep people first. OpenAI. https://cdn.openai.com/pdf/561e7512-253e-424b-9734-ef4098440601/Industrial%20Policy%20for%20the%20Intelligence%20Age.pdf
- World Economic Forum. (2026, April 9). Leaders are moving from systems of record to systems of work. World Economic Forum. https://www.weforum.org/stories/2026/04/how-leaders-using-ai-are-moving-from-systems-of-record-to-systems-of-work/