AI Threatens ‘Jobs With Dignity’, Warns Nobel Economist Simon Johnson: A Wake-Up Call for European Policymakers
Simon Johnson, the MIT economist who shared the 2024 Nobel Prize in Economic Sciences with Daron Acemoglu and James Robinson, warned in May 2026 that artificial intelligence threatens jobs with dignity — a class of middle-class employment that has historically anchored economic mobility and democratic stability. The warning, delivered at a series of conferences across Europe and the United States, is framed as a wake-up call for policymakers who, in Johnson’s view, are failing to grasp the scale and speed of the transformation.
The argument
Johnson’s central thesis builds on his Power and Progress work with Acemoglu (2023): technology choices are political choices. There is no automatic relationship between productivity gains from new technology and shared prosperity for workers. Whether AI delivers broad benefits or concentrated wealth depends on policy design: how AI is regulated, who controls AI infrastructure, how labour-market institutions adjust, how taxation and public investment respond. Without active policy choices, the default trajectory is one of job destruction without compensating creation of equivalent dignity-bearing employment.
What ‘jobs with dignity’ means
Johnson defines jobs with dignity as employment that combines reasonable economic security, meaningful task content, recognised social status and capacity for self-determination at work. These have historically included middle-class clerical work, paralegal services, customer support, mid-level finance and accounting roles, journalism, teaching, parts of healthcare administration. Generative AI, in Johnson’s analysis, is now displacing these roles — not by replacing humans entirely, but by hollowing out the cognitive and judgement-based content that gave them meaning.
The middle-class erosion thesis
Johnson’s empirical claim is that the middle of the labour market is contracting fastest. Routine cognitive work — historically the foundation of stable middle-class incomes — is the most exposed to AI substitution. At the bottom of the income distribution, manual and care work remain difficult to automate. At the top, highly specialised and creative work continues to command premium returns. The result is an hourglass economy where the middle is squeezed, with profound consequences for political stability and the social contract.
The European angle
For Europe, Johnson’s warning resonates with current debates around the AI Act, the proposed Algorithmic Management Directive, and the broader European Semester economic governance framework. Europe has historically maintained stronger labour-market institutions than the United States — collective bargaining, codetermination, social protection. These institutions could provide more capacity to manage the AI transition equitably, but only if European policymakers explicitly defend and adapt them. Johnson points to Germany, the Nordic countries and the Netherlands as examples of how social-partner dialogue can shape technology adoption.
Policy recommendations
Johnson outlines five priority interventions. First, active labour-market policies with substantially increased public investment in lifelong learning and retraining. Second, algorithmic management regulation — limiting the capacity of AI systems to monitor, evaluate and discipline workers without human accountability. Third, tax reform that does not privilege capital substitution of labour. Fourth, strengthened collective bargaining, including for platform and gig workers. Fifth, strategic public investment in complementary technologies — those that augment rather than replace human work.
The risk of inaction
Johnson is explicit about the political risk of policy inertia. “If middle-class jobs erode without an alternative, working people will turn against the political and economic system that failed them,” he argues. The election of populist and anti-system political forces across the West is, in his analysis, partly a consequence of the previous wave of automation and globalisation that displaced manufacturing employment. The AI wave, if mismanaged, will be larger in scope, affecting more sectors and more occupations simultaneously.
An opportunity, not a destiny
Johnson is careful to insist that the trajectory is not inevitable. AI can be deployed to augment human work, expanding the scope of meaningful employment, increasing productivity and freeing time for high-value activities. But this requires deliberate policy: “AI is not destiny. It is a choice.” European policymakers, he concludes, have perhaps two to three years to establish the institutional framework that will determine whether the AI transformation strengthens or erodes Europe’s social model. After that, path dependence will set in — and reversing course will become exponentially harder.
