How Progress Ends: Technology, Innovation, and the Fate of Nations (2025)

By Carl Benedikt Frey

Carl Benedikt Frey is a Swedish-German economist and economic historian. He is currently the Dieter Schwarz Associate Professor of AI & Work at the Oxford Internet Institute and a Fellow of Mansfield College, University of Oxford. He is widely known for his joint work with Michael A. Osborne in their 2013 paper, ‘The Future of Employment,’ which estimated that 47% of US jobs were at high risk of automation, becoming one of the first and most cited works on AI's impact on modern labour markets. Frey is also the author of The Technology Trap (2019) and How Progress Ends (2025), which was a finalist for the Lionel Gelber Prize and the Financial Times Business Book of the Year Award. From his background, Frey has the academic pedigree to make his central argument both convincing and compelling.

In How Progress Ends, Carl Benedikt Frey offers a sobering corrective to the complacent assumption that technological advances naturally deliver sustained economic growth and social progress. Drawing on economic history, political economy, and contemporary debates about artificial intelligence, Frey consistently argues that what we define as progress is neither automatic nor irreversible. Instead, it depends on fragile institutional, political, and social arrangements that can — and often do — break down, especially when put under continuous strain.

The book’s central claim is historical rather than technological. Frey contends that societies frequently mistake bursts of innovation for permanent trajectories of growth. By revisiting episodes such as the late Roman Empire, Imperial China, early modern Europe, and the Industrial Revolution, the book shows that technological capability alone has never guaranteed prosperity. Political and economic elites can block diffusion of wealth, labour-saving technologies can deepen inequality, and political coalitions can undermine the incentives required for sustained innovation. Progress ends, Frey argues, when institutions cease to align innovation with broad-based economic inclusion.

One of the book’s strengths is its refusal to frame current debates about AI as unprecedented. Frey situates contemporary anxieties within a long lineage of technological disruption, reminding readers that automation has repeatedly threatened employment, social stability, and political legitimacy. This historical grounding usefully tempers both techno-optimism and the apocalyptic rhetoric that comprise much of our contemporary discourse about AI. The discussion of how societies previously adapted — or failed to adapt — to general-purpose technologies is among the book’s most compelling contributions.

Equally persuasive is Frey’s emphasis on political economy. Rather than treating innovation as a neutral force, the book highlights how power, rents, and institutional capture shape technological outcomes. When political and economic elites benefit from stagnation, or when the losers from change are politically marginalised, progress can stall or reverse. This framework is especially effective in analysing contemporary economies marked by rising concentration of wealth, declining labour bargaining power over wage growth, and uneven or declining productivity growth.

That said, the book occasionally risks overstretching its historical analogies. While the breadth of perspective is illuminating, some comparisons — particularly between pre-industrial societies and modern, globally integrated economies — can feel more suggestive than decisive. Readers may also find that the policy prescriptions, though sensible, remain relatively high-level. Frey is significantly clearer about the dangers of institutional decay than he is about the concrete political pathways required to avoid it.

One of the chief merits of How Progress Ends is how directly it speaks to contemporary policy debates that are often framed as technical rather than political. Frey’s historical argument — that progress depends on institutions which distribute the gains from innovation widely enough to sustain political legitimacy — cuts across several of today’s most contested policy areas. Relating How Progress Ends to current policy debates, there are eight factors which need to be considered in turn:

First, the book challenges prevailing assumptions in AI and industrial policy.

Current debates in the UK, EU, and US often focus on accelerating innovation through deregulation, public–private partnerships, and strategic investment in frontier technologies. Frey’s analysis suggests that this focus is incomplete. Innovation-led growth can falter not because technologies fail, but because their gains are captured by narrow groups. This has clear implications for AI governance: policies that prioritise speed and global competitiveness without addressing labour displacement, market concentration, or diffusion risk reproducing the very dynamics that historically caused progress to stall.

Second, Frey’s emphasis on distribution speaks directly to productivity and labour-market policy.

Many advanced economies face the paradox of rapid technological capability alongside stagnant wages and weak productivity growth. Frey’s historical cases suggest that when workers lack bargaining power or access to complementary skills, labour-saving technologies reduce political support for innovation itself. This reframes debates over skills, collective bargaining, and welfare not as social add-ons, but as core components of a functioning innovation system.

Third, the book offers a critique of contemporary ‘growth-first’ strategies.

Policymakers often assume that growth will eventually resolve inequality, regional imbalance, and political disaffection. Frey reverses this logic: without institutions that embed innovation in inclusive economic structures, growth itself becomes fragile. This perspective resonates strongly with debates over regional development, infrastructure investment, and ‘levelling up,’ where technological adoption has often reinforced, rather than reduced, spatial inequality.

Fourth, Frey’s analysis bears on concerns about democratic backsliding.

By showing how economic exclusion and technological disruption have historically undermined political stability, the book connects innovation policy to broader questions of democratic resilience. If progress is perceived as benefiting only political and economic elites, societies may turn against openness, competition, and even technological advance itself – an increasingly visible risk in contemporary populist movements.

Taken together, How Progress Ends argues that innovation policy cannot be separated from institutional reform. The book implicitly critiques a policy discourse that treats technology as a substitute for politics. Instead, Frey insists that sustaining progress requires deliberate choices about competition policy, labour institutions, education, and the distribution of economic power. In this sense, the book is less a warning about technological limits than about political complacency.

Carl Benedikt Frey’s How Progress Ends is particularly resonant in the UK context, where policy discourse remains strongly committed to technological dynamism while struggling to reconcile innovation with weak productivity growth, regional inequality, and political disaffection. Frey’s central warning — that progress fails when institutions allow the gains from innovation to concentrate too narrowly — casts a critical light on several pillars of current UK policy.

Fifth, AI, regulation, and the UK’s ‘pro innovation’ stance.

The UK government has positioned itself as a global leader in ‘pro innovation’ AI regulation, emphasising flexibility, light-touch oversight, and rapid deployment. Frey’s historical argument suggests that this framing is incomplete. Innovation-friendly environments can still produce stagnation if new technologies reinforce monopoly power or displace labour without adequate institutional adjustment.

In UK terms, the risk is not over regulation but under institutionalisation: AI adoption that boosts firm level efficiency without translating into economy wide productivity gains or rising wages. Frey’s work implies that AI policy should be evaluated not only by its impact on growth or competitiveness, but by whether it strengthens diffusion, competition, and worker complementarity — areas where UK policy remains thin.

Sixth, productivity, skills, and the limits of supply side reform.

The UK’s long running productivity puzzle looms large over Frey’s analysis. Successive governments have treated productivity as a technical problem to be solved through skills, infrastructure, and R&D spending. While Frey does not reject these tools, How Progress Ends suggests why they have delivered disappointing results.

Historically, productivity growth falters when workers lack bargaining power or access to the gains from innovation. In the UK, weak wage growth, declining trade union membership, and high labour force turnover limit incentives for firms to invest in productivity enhancing technologies. Frey’s framework reframes debates over skills policy, collective bargaining, and in work security as integral to innovation itself, rather than as distributive afterthoughts.

Seventh, Industrial strategy and regional inequality.

Frey’s emphasis on institutional alignment speaks directly to the UK’s uneven economic geography. The concentration of innovation, capital, and high productivity firms in London and the Southeast mirrors historical cases where progress became spatially and socially narrow, ultimately undermining political support.

Policies branded as ‘levelling up’ or industrial strategy often focus on place based investment without addressing the deeper institutional asymmetries — planning constraints, finance allocation, governance capacity — that shape where innovation actually takes root. Frey’s historical lens suggests that without tackling these structural factors, regional policy risks becoming symbolic rather than transformative.

Eighth, democracy, legitimacy, and technological change.

Perhaps most pointedly, How Progress Ends speaks to the UK’s post Brexit political economy. Frey shows that when large segments of society experience technological change as exclusionary or destabilising, they may turn against openness, competition, and expertise itself. The UK’s fractured political landscape, declining trust in institutions, and scepticism towards elite driven economic narratives echo this pattern.

From this perspective, innovation policy is inseparable from democratic legitimacy. A strategy that celebrates technological leadership while tolerating stagnant living standards and insecure work risks reproducing the very conditions under which progress historically stalled.

Nonetheless, How Progress Ends is a timely and intellectually serious intervention. It challenges the comforting belief that innovation will inevitably rescue stagnating economies or fractured societies. Instead, it insists that progress is a political achievement — one that must be actively defended and renewed. In an era of rapid technological change and growing social tension, Frey’s warning deserves careful attention, especially when read through a UK policy lens, where How Progress Ends is less a critique of technology than of complacency. It challenges the assumption that innovation can substitute for institutional reform, or that growth alone will repair social fractures. For the UK, Frey’s message is clear: sustaining progress requires rebuilding the political and economic institutions that ensure technological change delivers visible, shared benefits. Without that alignment, even a highly innovative economy like the UK’s can drift into stagnation.

Reveiewed by Maj Jason Fensome

Published by: Princeton University Press, 16 Sept. 2025.

552 pages. ISBN-13: 978-0691233079