AI is widening the divide between workers who use it and those managed by it

AI is creating a two-tier workforce

AI worker surveillance is already reshaping working conditions across Britain and beyond, creating a stark divide between those who use artificial intelligence as a tool and those who are governed by it. In warehouses, delivery networks, and call centres, AI-powered systems now dictate shift assignments, measure task completion times, and grade performance on metrics that workers cannot see or contest.

For higher-paid professionals, analysts, lawyers, consultants, and managers, AI often functions as a copilot. It removes repetitive tasks, accelerates research, and creates space for more complex thinking. But for workers in logistics, retail, hospitality, and the gig economy, the experience is entirely different. AI is not an assistant. It is a mechanism of control.

Research conducted over the past decade on worker and AI coexistence, cited in the 2024 US White House economic report, suggests the most urgent question about AI's impact on work is not imminent mass unemployment. It is the growing gap in skills, autonomy, and wellbeing between those who work alongside AI and those who find themselves managed by it. Tezons's coverage of tech layoffs and AI investment uncertainty examined whether productivity gains can justify the current scale of automation, a question that sits at the heart of the business case for AI deployment.

Bossware is already widespread in UK workplaces

A third of UK employers currently use so-called bossware technology to monitor workers' online activity. Scheduling tools, route optimisation software, and automated performance dashboards are among the systems being deployed to determine who gets a shift, how long each task should take, and whether an individual is operating at maximum capacity.

The spread of these tools is not confined to low-wage sectors. Amazon's software engineers have reported being surveilled and pressured to use AI to hit higher productivity targets, even when doing so demonstrably slows their work. Meta has announced plans to track employee keystrokes, mouse movements, and clicks to train its AI models. Workers who currently benefit from AI assistance may find themselves subject to similar oversight as these practices become normalised across corporate structures.

For those already subject to algorithmic management, the consequence is stress. When every movement, pause, and decision can be measured and rated by an opaque system that offers no meaningful avenue for challenge, the psychological burden is significant. Work is not only a source of income. It is bound up with dignity, trust, and the degree of control a person retains over their own working day.

Training investment is failing to reach the workers who need it most

A global survey of business leaders found that whilst most acknowledge AI skills as a source of competitive advantage, relatively few have committed meaningful budgets to employee AI training. Governance frameworks are sparse, and many line managers carry little formal responsibility for helping their teams adapt to AI-transformed roles.

This creates a structural problem. Better-paid workers with greater job autonomy receive training and tools that extend their capabilities. Lower-paid workers, more likely to be subject to algorithmic management, receive neither the training nor the protections that would allow them to adapt. The gap between the two groups is not closing.

Britain has announced programmes to expand AI skills across the workforce, with ambitions to reach millions of adults through subsidised online learning. These efforts address part of the problem. They do not address the governance failure: organisations are deploying AI systems that affect pay, performance ratings, and job security without adequate transparency or accountability. A survey of business leaders found that even where AI training budgets exist, lower-paid and part-time workers are disproportionately excluded from access.

What AI workplace surveillance means for UK workers and employers

The pattern of algorithmic management visible in warehouses and delivery networks is not confined to those sectors for long. The same methods being refined in logistics are expected to spread to offices, hospitals, and schools as AI tools become cheaper and more capable. Workers who currently sit on the empowered side of the divide should not assume their position is fixed.

Researchers argue that meaningful reform requires two things. First, workers must have a voice in how AI systems are introduced in their workplaces. Studies show that involving workers in the design and rollout of AI tools produces better outcomes for both job quality and business performance. Second, systems that affect pay and performance must be transparent and open to challenge. Algorithmic decisions presented as neutral or objective are neither, and workers need clear routes to contest outcomes that affect their livelihoods.

The choices determining which workers benefit from AI and which are subjected to closer control are being made employer by employer, across Britain, right now. Without effective governance and a genuine commitment to worker involvement, the AI workplace divide will embed itself as quietly and as durably as any previous form of structural inequality in the labour market.