In recent years, the notion of “progress” has become something of a commonplace in UK policy circles. You're either for progress, or you're against it. And no one wants to be against progress. The UK's “progress movement”—a loose coalition of organisations ranging from the Centre for British Progress on the centre-left to outlets such as Works in Progress, edited by former Adam Smith Institute chief Sam Bowman—is often associated with the American movement for “abundance”, popularised by Ezra Klein and Derek Thompson's book of the same name.1 Yet it is animated by its own characteristically British bogeymen—from crumbling Victorian-era infrastructure to Home Counties NIMBYism—and nostalgia for a bygone era in which the British state possessed the “capacity to marshal material resources towards ambitious common purposes”.2 It is hardly surprising that progress has become such a prominent motif in the UK of 2025. Last year IPPR found that we have had the lowest level of investment in the G7 for 24 of the last 30 years; state capacity has been emaciated by privatisation and outsourcing; our post-crash “lost decade” of productivity and GDP growth is now well into its teenage years.3 What is more striking, however, is what these invocations of progress are doing. It seems to me that visions of future abundance often function less as promises to be redeemed than as a kind of ultimatum, a set of demands to be answered. This dynamic is particularly evident in debates around artificial intelligence (AI), a set of technologies that in recent years has been cast by proponents as central to the future prosperity of the UK and indeed the planet.4 It's not uncommon to hear, for instance, that the UK government needs to prioritise energy-intensive datacentre infrastructure for much of our grid capacity and remaining carbon budgets, and that to do otherwise would be to jeopardise our status as an “AI leader”.5 Or that our intellectual property regime needs urgent reform, because requiring tech companies to license assets from creatives would sound the death knell of the AI sector.6 “Trade-offs” of this sort are frequently invoked in AI policy, but they are too often framed as if they are inevitable costs rather than choices to navigate. The unspoken implication is that the investments and regulatory changes demanded by large technology companies are the only solution and to do anything else is to strangle growth. If you're against them then you're against progress. The trolley is hurtling down the tracks, the destination is set: forget about flicking a switch to divert course, let alone pulling the brakes. The ultimatum is not an uncommon feature of politics. “You're either with us or you're against us” is an effective rhetorical cudgel, as is its Thatcherite sibling “There is no alternative”. Progress discourse does, however, have certain features that foreclose debate and, in so doing, make it more difficult to talk meaningfully about the future. In the first place, there is something peculiar about the idea that one can be for or against “progress” tout court. “Technology” is not one thing: an electric car is not a chatbot is not a cancer drug. Nor can the futures augured by technological progress be equated: a world of ubiquitous private electric vehicle use is manifestly different to one in which public transport is cheap and abundant. Conceiving of progress as a unitary phenomenon occludes these differences, preventing us from articulating which we prefer. All is collapsed into one: particular technologies, and the interests attached to them, are rendered synonymous with a universal and public good. This sleight-of-hand trick has implications for how we think about policy. It casts complex socio-political challenges as engineering problems fit for technical optimisation. There are “pro-growth policies” and “anti-growth coalitions” that can be assessed according to their notional impact on productivity and GDP.7 Obstacles to development—whether trade union and community mobilisation, regulatory action, or political intervention—can only be understood as costly vetoes imposed by the few on the many. Curtailing veto power is therefore a signal aim of many in the progress movement. Proposals to exempt datacentre projects from planning approvals, for example, are justified on the grounds that they would allow the common good of development to be driven forward without the interference of partial or vested interests.8 Our notions of “common good”, however, are invariably figured by and for other vested interests with capital to mobilise for development and the political will to enforce a vision by fiat. Once this is understood, the presence of worker, community and regulatory veto power looks less the imposition of a cost and more like the demand to simply be allowed into the room. I am sympathetic to many of the substantive demands of the progress camp. Few, I think, would defend our existing planning system in its totality; greater investment in science, technology and infrastructure is sorely needed. I am, however, allergic to the politics of the ultimatum. It cedes collective responsibility for determining our future, seeing it as something to be outsourced to technical experts, market analysts, and those actors with the deepest pockets. In the words of political theorist Wendy Brown, “public life is reduced to problem solving and program implementation, a casting that brackets or eliminates politics, conflict, and deliberation about common values or ends”.9 It is true that technological revolutions have typically proceeded by means of unilateral imposition rather than consensual negotiation. There were no stakeholder roundtables when the factory-owners of Lancashire introduced the spinning jenny; Henry Ford's rollout of the Model T afforded little space for “deliberation about common values or ends”. Nonetheless, history also shows us that contestation of technological development from below can direct innovation down different paths. The rapid industrialisation of England during the eighteenth and nineteenth centuries provides a paradigmatic example. Campaigns against inhumane working conditions led to the Factory Acts, the establishment of quasi-regulators, and—eventually—the formalisation of the trade union movement.10 This dialectic between regulatory reform and grassroots action gave working people some leverage over the bruising transformations upturning their lives. Over time, these struggles not only changed the conditions under which technology was used but reshaped how those same technologies were designed, leading to fundamentally different forms of factory organisation. What this illustrates is that technological development is underdetermined, to borrow a term from philosopher of technology Andrew Feenberg.11 Technical factors shape the form and function of new technologies, but never completely. An internal combustion engine has certain technical specifications that need to be met if it is to power an automobile, but not every aspect of a car is determined by those specifications. There is no technical reason an automobile should look like the cars of today, but social convention and regulatory requirements make it so. This suggests that technology can be developed in different ways: there are always other paths innovation can take. The point I am making is not that we should envy the conditions of earlier technological revolutions, of the Lancashire mill workers or the early industrial proletariat. It is rather that the direction of progress is closely connected to the presence of rival sources of power—from the regulatory state to the workers' movement—that can press those developing and deploying technology to adopt a wider perspective. The technological foundations of modern Britain owe as much to these forces as they do to the brilliance of past innovators and the largesse of industrialists grown rich on imperial wealth. After decades of hollowing out, however, our traditional bases of public and countervailing power no longer seem adequate to the task of contesting technological change.12 In the UK we have historically low levels of unionisation in most sectors, entrenched by successive waves of anti-union legislation that are only now beginning to be unwound.13 Despite some limited uplifts in funding and improved coordination efforts, regulators are often insufficiently resourced to manage the challenges posed by new technologies, and increasingly face obligations and ministerial directives that co-opt them to the growth agenda at the expense of their protective duties.14 In the absence of these forces, and without recourse to more holistic avenues for collective deliberation about the future, ordinary people have been left with scant options for reclaiming agency over technology. For many, this stokes pervasive feelings of “digital resignation”: a sense that not all is right with our technological environment, but that alternatives are scarce or impossible to pursue.15 Others are turning towards legal obstruction and direct action to reassert some semblance of control.16 Not everyone, it turns out, responds well to a threat: deprived of the opportunity for negotiation, some switch off, while others reach for the veto. AI is no exception to this story of technological underdetermination. Social and economic factors—who owns it, who deploys it, who is subject to its use—continue to shape how AI is experienced by people across the world. In the contemporary context, this means that a small number of firms enjoy a de facto monopoly over critical AI infrastructure and therefore the agenda for future research and deployment.17 As the AI Now Institute's Kate Brennan, Amba Kak, and Dr. Sarah Myers West have articulated, “AI as a field has been not just co-opted but constituted by the logics of a few dominant tech firms”.18 The corollary of this is that many other constitutions are possible. “AI” does not need to mean continuing to stoke a hyperscale paradigm designed to ensure power and capital continue to flow to a small number of companies on the west coast of the United States. Unwinding the concentration that exists in today's AI sector could unlock a plurality of different pathways for development, a kaleidoscope of AI futures. The evidence suggests that people up and down the country understand this: they recognise that there is no singular “AI revolution” that is going to work for everyone and want more autonomy and control over how AI manifests in their own lives.19 As Eleanor Shearer and I have previously written, a progressive strategy for opening up these possibilities would need to comprise two elements.20 There is, firstly, a need to strengthen “downstream” regulatory protections, strengthening the veto power of trade unions and other actors capable of mobilising on behalf of affected groups. Proposals set out by the Trades Union Congress would provide unions with rights to engage over technology deployed in their workplaces, while the Welsh Labour government has introduced a social partnership approach for algorithmic management which promises a say for workers in how these systems are used.21 Powers such as this need to be seen not just as obstructive but as prefigurative, starting points for the constructive negotiation needed to tailor AI implementation to diverse needs and perspectives. In isolation, however, a platform for building worker and community influence over deployment is likely to founder against the agenda-shaping power of the technology companies. This points to the second component of any progressive AI strategy, which is to rein in that power and diversify “upstream” innovation. Public investments in AI should be geared towards pluralising AI research and building independent capacity.22 Structural measures should be introduced to promote competition, reduce hoarding of talent and intellectual property, and restrict predatory corporate partnerships.23 The regulators and the administrative state ought to be bolstered with the requisite capacity to steer innovation without relying solely on industry expertise for the identification of “opportunities” and “risks”.24 The possibility of a more plural AI future matters beyond the relatively-narrow confines of industrial policy and competition economics. As well as being indelibly shaped by social factors, AI has the potential to shape our societies in turn, creating unintended and profound ripple effects. The introduction of live facial recognition technologies in public places, the embedding of foundation models in public services, the establishment of platforms that allow AI companies to access public data: these are sweeping changes that threaten to fundamentally alter the social contract between citizens, private companies, and the state. The recent “deliberative turn” in technology policy has emphasised the need for greater participation from workers, service users, and affected communities in managing such changes. The history of innovation shows us, however, that meaningful technological choices can only be underwritten by the right to choose differently, push back, and ultimately say “no”. Veto power is—painfully, frustratingly, maddeningly—the basis of democracy. Once another person, or group of people, has a stake—and the right of refusal—then the issue at hand ceases to be a problem that can be fixed unilaterally. By trying to eliminate the veto power that gives people a stake in decisions, many in the progress movement betray a view of democratic participation merely as friction to be eliminated rather than a source of earned legitimacy. Progressives would do better to flip the equation: to see friction as an important wellspring of collective intelligence. When trade unions resist the introduction of an algorithmic tool, when data workers blow the whistle about unacceptable conditions, when communities push back against the impact of a proposed datacentre, these are not obstacles to be managed away but signals about the technological future that people want to see. The route to better and faster development is not to ignore these signals but to plan and mobilise investment in a way that takes them into account. Certain corners of the progress movement understand this. Within the ranks of those who advocate for “progress” are those who recognise that their desired ends may be aided by measures such as greater devolution, micro-democratic interventions such as “street votes”, and investment in greater regulatory and state capacity to facilitate trusted deployment.25 There may be sufficient overlap to allow for common cause to be made with progressives advocating for greater public and community power in and over technology. Citizens' assemblies and other deliberative fora; social partnership approaches; regulatory and government capacity to make better decisions about technology without relying solely on industry expertise; trade unions and civil society organisations empowered to demand better futures, including to mobilise for development if it is the right thing. We need more of all of this. But we also need to move past the idea that there is some sort of neat policy fix that will eliminate conflict over technology and allow for stable and congenial deliberation. We need to become comfortable with the idea that the road to progress is never a smooth one, and that technological change demands struggle. Above all what we need is to democratise the future: to move from “There is no alternative” back to “What do we want?” This means abandoning the politics of the ultimatum and returning to the agonising and agonistic work of collective deliberation about our underdetermined technological choices. This work will take us beyond technology policy and into the unruly realm of the political—but this, of course, was the world we were inhabiting all along. The author declares no conflicts of interest.
Building similarity graph...
Analyzing shared references across papers
Loading...
Matt Davies
IPPR Progressive Review
Economic Policy Institute
Building similarity graph...
Analyzing shared references across papers
Loading...
Matt Davies (Mon,) studied this question.
www.synapsesocial.com/papers/69b3aaa802a1e69014ccb7a4 — DOI: https://doi.org/10.1111/newe.70008