The deployment of artificial intelligence in the workplace has introduced a new and structurally distinct challenge to collective attention. Previous disruptions to working attention — the open-plan office, the smartphone, the always-on email norm — were largely environmental: they created conditions in which attention was more easily interrupted or claimed by others. AI systems at work operate differently. They do not merely interrupt attention; they reshape the landscape of what work requires attention for, displace certain cognitive tasks while intensifying others, and introduce a new layer of monitoring, optimization, and task-assignment that operates beneath and around the worker's conscious management of their own focus.
The most visible AI applications in contemporary workplaces include automated scheduling and task-prioritization tools, AI-generated communication drafts, algorithmic performance monitoring, intelligent meeting summarization, and large language model assistants integrated into productivity software. Each of these interventions has a complex and often contradictory relationship with worker attention. AI-generated email drafts, for instance, may reduce the cognitive load of routine correspondence — freeing attention for more complex work — but simultaneously accelerate the rate at which correspondence is generated and expected, increasing the total volume of communications demanding response. The net attentional effect may be negative even when the per-message cognitive cost is reduced.
Algorithmic management — the use of AI and data systems to direct, monitor, and evaluate worker behavior in real time — represents a more fundamental reorganization of the attention landscape. In warehouse and logistics settings, workers receive continuous task assignments via handheld devices or earpieces, with response times monitored to the second and performance scores calculated algorithmically. In customer service, AI systems monitor call scripts, flag deviations from protocol, and score emotional tone in real time. In professional settings, productivity monitoring software tracks keystrokes, application usage, active versus idle time, and communication frequency. This monitoring creates a form of continuous attentional demand that is qualitatively different from the traditional supervisor relationship: the algorithmic gaze is ceaseless, granular, and impersonal, and workers report that awareness of it produces a form of sustained low-grade vigilance that is cognitively and psychologically costly.
The attentional consequences of AI deployment at work also include the emergence of what might be called cognitive deskilling — the progressive withdrawal of complex cognitive challenges from workers whose tasks are increasingly automated, leaving them with the monitoring and exception-handling residual that AI systems cannot yet manage. This residual is cognitively demanding in a specific way: it requires sustained vigilance for rare events amid long periods of routine — the "vigilance decrement" problem that psychologists have studied in radar operators and air traffic controllers. A workforce managed by AI into patterns of interrupted deep attention and sporadic exception-response is a workforce whose collective cognitive capacity is being restructured in ways that may be difficult to reverse.
At the collective level, the deployment of AI in workplaces raises governance questions that no existing labor law framework was designed to address. Traditional working time law regulates clock hours; it cannot regulate the attentional intensity of those hours. Traditional occupational health law addresses discrete physical and psychological hazards; it struggles to address diffuse systemic changes in the cognitive environment of work. Trade unions, where they exist, are beginning to negotiate AI governance provisions into collective agreements — provisions covering transparency about monitoring, limits on algorithmic management, and worker rights to explanation of AI-generated performance assessments. These negotiations are at an early stage and their outcomes are uneven.
The deeper question that AI at work raises for collective attention is a distributional one. When AI systems automate cognitive tasks, the value of the cognitive labor that remains — the creative, relational, and judgment-intensive work that AI cannot yet replicate — is likely to rise. But access to the conditions that develop and sustain this higher-order cognitive capacity is unequally distributed. Workers with more education, more autonomy, and more favorable working conditions are better positioned to cultivate the attention management skills, the creative capacity, and the metacognitive awareness that the AI economy is beginning to reward. Workers whose tasks are most fully subject to algorithmic management are least positioned to develop these capacities. The AI transition at work may thus produce a new form of attentional inequality: a polarized landscape in which some workers have their higher-order cognitive capacities expanded by AI tools while others have their attention reduced to the monitoring of automated systems.