Work & Contribution
A weak safety net and virtually no social partnership: transitions in AI-driven office work are poorly cushioned.
Where does the USA stand, and where does it break?
The United States has one of the most flexible labour markets among industrialised countries. Employment is predominantly at-will — in principle without statutory protection against dismissal and without the notice periods common in Europe. Trade union membership stands at only around 10 per cent (U.S. Bureau of Labor Statistics, 2024), and considerably less in the private sector. There is no nationwide system of social partnership or short-time work schemes in the European sense. Only individual states offer limited short-time compensation programmes. The labour market organises adjustment through mobility rather than protection: workers change jobs frequently, often several times over their working lives, and gig and project-based work are widespread. This channels displaced workers swiftly into new roles, but leaves them without support during the transition. The labour force participation rate is around 62 to 63 per cent (BLS). Unemployment insurance is a state-level matter, typically capped at 26 weeks and offering a low wage-replacement rate. Active labour market policy and publicly funded retraining are underdeveloped by European standards. Retraining is largely market-based and privately funded.
Generative AI hits well-paid cognitive office roles first: legal assistance, accounting, customer service, programming, and basic analysis. In a labour market that organises adjustment through rapid reallocation, this can mean either a swift return to new roles or a hard fall with no safety net. Because short-time work schemes, social partnership, and publicly funded retraining are absent, workers bear the risk of transition largely on their own. The more highly qualified and specialised the displaced role, the steeper the potential fall. High job mobility is both a strength and a weakness: it accelerates structural change, but leaves the costs with the individual. Whether AI strengthens the American middle class by making workers more productive, or hollows it out by displacing entire occupations, will be determined by access to retraining, adequate income protection, and capital income as a second financial foothold.
Paths for the USA
The United States decouples social benefits from individual employers and ties them to the worker, complemented by wage insurance that partially offsets income losses when workers move into lower-paid employment. This fits the flexible, mobile labour market and the prevalence of gig work, as protection follows the person rather than the job. A number of states and platforms are already experimenting with portable models. The drawback: implementation and funding are complex, and wage insurance cushions the fall rather than preserving employment.
Trade-off: flexibility with a safety net versus administrative complexity
Feasibility: The approach most aligned with US labour market culture
The federal government broadens the short-time compensation programmes currently available only in selected states, under which workers receive partial wage replacement and reduced hours instead of being laid off. This approach has proved an effective buffer in Europe, keeping workers attached to their employers. In the United States, however, it is weakly embedded, rarely used, and culturally unfamiliar, as employers value the ability to reduce their workforce quickly. Where entire occupations face structural displacement, short-time working delays change rather than shaping it.
Trade-off: employment stability versus speed of adjustment
Feasibility: Better suited to cyclical downturns than to structural breaks
The state guarantees publicly funded work in areas of social need — care, infrastructure, education, and the environment — when the market fails to provide sufficient jobs. This secures employment and purpose when AI displaces private office work, while also strengthening undersupplied sectors. In the United States, a Job Guarantee faces deep opposition rooted in free-market regulatory principles, is regarded as government overreach, and would be costly. It also competes with private employment and is difficult to tailor to the skills of displaced knowledge workers.
Trade-off: employment security versus market logic and cost
Feasibility: Politically difficult to advance within the United States' market-oriented culture
The United States expands publicly funded retraining, career counselling, and job placement. This includes individual training accounts, sectoral partnerships with employers, and an entitlement to reskilling. It addresses the decisive point: in a flexible labour market, the transition matters, not protection against dismissal. Experience with sectoral training programmes shows measurable results. The weakness: the existing framework (WIOA) is underfunded and fragmented, and retraining only succeeds when new roles actually emerge.
Trade-off: empowerment versus the cost and effort of funding and coordination
Feasibility: Sectoral partnerships are the most effective component
Trigger-based phases
Trigger: Employment in individual exposed office occupations declines measurably as new hiring falls back, rather than triggering mass redundancies.
The flexible labour market channels workers into new roles, but transitions are unprotected. Initial sectoral retraining programmes and portable benefit models are being piloted.
Trigger: Entire knowledge-work occupational profiles shrink permanently, and displaced professionals search longer for work despite flexible labour markets.
The limits of pure mobility become apparent. Calls for wage insurance, portable benefits, and publicly funded retraining grow louder. States and the federal government compete to provide answers.
Trigger: A lasting reform combines market flexibility with portable benefits and a reliable entitlement to reskilling.
The United States must decide whether its flexible labour market acquires a safety net capable of supporting transitions, or whether it remains dynamic but unprotected. The answer will determine whether AI strengthens or hollows out the middle class.
Real-world pilot programmes
Status: In force in roughly half of all states, but little used
Under Short-Time Compensation, employees whose hours are reduced receive a proportional wage supplement instead of being made redundant. The programme illustrates two things: a short-time working mechanism exists in US law, and its uptake depends heavily on how firmly it is embedded and what incentives are in place.
Status: Active (federally funded retraining in partnership with employers)
Under the Workforce Innovation and Opportunity Act, federal and state governments fund retraining tailored to specific roles in partnership with employers. Successful sectoral programmes demonstrate measurable income gains and serve as a model for AI-driven transitions.
Status: programme largely wound down in 2022; remains relevant as a case study
This decades-long programme supported trade-displaced workers with retraining and income assistance. Its expiry and mixed track record offer valuable lessons for how an effective transition programme for AI-displaced workers should be designed.
What we are watching
This metric measures whether AI is genuinely displacing well-paid knowledge work. The threshold matters because a sustained decline strikes at the core of the American middle class. A tipping point would signal that displacement is shifting from the exception to a structural feature. For policymakers, it is the sharpest indicator of whether the flexible labour market can still absorb transitions.
This measures how quickly the labour market reabsorbs displaced workers. The metric matters because longer search times coincide with short, inadequate income support, causing people to fall through the net. If it tips, mobility ceases to function as a buffer. Politically, this forces a debate about the duration and level of income support.
This metric shows whether the United States is keeping retraining in step with the pace of automation. It matters because, in a flexible labour market, retraining determines access to new occupations. An upward inflection would demonstrate that transitions are succeeding. If participation remains low, displacement becomes entrenched.