Marx observed that machinery under capitalism serves a dual function: it increases the productivity of labour, and it disciplines labour, making workers more replaceable, more controllable, and more dependent on the capitalist who owns the machinery.

AI performs both functions simultaneously, and at a scale Marx could not have anticipated.

Automation and displacement. The debate about AI and employment has been ongoing since at least 2013, when Frey and Osborne published research suggesting that 47% of US jobs were at high risk of automation. Subsequent research has produced more nuanced estimates, but the direction is clear: AI is automating cognitive tasks that previously required human workers, across a wide range of sectors, legal research, financial analysis, medical diagnosis, customer service, content creation, software development.

Aaron Benanav’s analysis of labour market dynamics argues that the problem is not primarily technological unemployment, the displacement of workers faster than new jobs are created, but secular stagnation: a slowdown in economic growth that reduces labour demand across the economy, into which AI-driven displacement is accelerating. The result is not joblessness but labour market polarisation, a growing gap between high-skill, high-pay work and low-skill, low-pay work, with the middle hollowing out.

Algorithmic management. For workers who remain employed, AI increasingly determines the conditions of their work. Warehouse workers at Amazon have their every movement tracked by AI systems that set pace targets, monitor breaks, and flag underperformance for management review. Ride-share drivers are managed by algorithms that determine which rides they are offered, what rates they are paid, and when they are deactivated. Call centre workers have their conversations monitored, scored, and used to set performance standards.

This is what Marxist analysts call the intensification of labour, using technology to extract more work from each worker-hour. It is also a form of surveillance that extends managerial control into domains previously outside it: the pace of movement through a warehouse, the routes taken between deliveries, the emotional tone of customer interactions.

The gig economy as AI-managed precarity. The so-called gig economy, Uber, Lyft, Deliveroo, TaskRabbit and hundreds of similar platforms, is an AI-managed labour market that systematically denies workers the employment protections won by labour movements over a century. Gig workers are classified as independent contractors, giving them flexibility, in the platform’s framing, or denying them benefits, job security, and labour rights, in the Marxist framing. The platforms’ AI systems manage these workers with the granularity of an employer while maintaining the legal fiction of independence.

Reflection question: How has AI changed the conditions of work in your sector? Who has benefited? Who has borne the costs?