Access Official Technical PDF

Share Summary:LinkedInEmail

Computerization and employment are often discussed in terms of jobs gained or lost. But technology can also change which tasks workers perform, how work is organized and how wages evolve. A task-based analysis of U.S. labor trends offers a useful perspective for automation engineers, plant managers and workforce planners: the effects of technology vary with both the autonomy a task requires and the level of social perceptiveness involved.

The analysis examines successive technology eras, from mainframes and personal computers to networked, mobile and Internet of Things systems. Its central implication is not that one group of workers is uniformly protected or displaced. Rather, computerization interacts with the characteristics of work, influencing employment and compensation in different ways.

Why analyze tasks, not just occupations?

An occupation usually combines multiple activities. Looking only at job titles or education levels can therefore obscure how technology affects the work itself. The study represents tasks along two dimensions: the degree to which work is structured, and the social perceptiveness it requires.

Structured work leaves less discretion to the worker over tasks, priorities and goals. At the other end of the scale, highly autonomous work involves more judgment and independence. Social perceptiveness refers to awareness of other people’s reactions and an understanding of why they respond as they do. These dimensions distinguish, for example, a tightly specified task with little interaction from work that requires both discretion and sustained engagement with other people.

This framework matters for automation because computerization can change the productivity of capital and labor, and the degree to which they can substitute for one another. Data processing and prediction may make some activities easier to automate, while tasks requiring human judgment or interaction can present different challenges. The paper treats these as patterns in work, rather than assuming that a single attribute such as education explains labor-market change.

Computerization and employment across technology eras

The paper groups technological change into eras associated with shifts in computing devices and connectivity. Mainframes supported organizational processes such as accounting and payroll. Personal computers brought computing to the desktop, while the network era connected systems and enabled wider collaboration and online business activity. Mobile platforms extended connectivity, followed by an IoT era characterized by connected devices and machine-to-machine communication.

These periods provide context for changes in workplace technology. A new platform can affect more than the speed of an existing process: it may alter how work is coordinated, which tasks are needed and how production is organized. The framework therefore considers technology as a source of structural change with uneven effects across tasks, rather than as a single, uniform increase in productivity.

For industrial organizations, this is a useful distinction when evaluating automation. A process may contain tasks that are repetitive and highly structured alongside activities that require troubleshooting, coordination or interaction. Assessing a job only as “automatable” or “not automatable” can miss those differences. A task-level view can help decision-makers ask which activities are changing and what kinds of human capabilities remain part of the work.

What the analysis finds about wages and employment

The paper reports several findings that complicate simple predictions about technology and work. High-autonomy tasks experienced the greatest wage growth, while tasks with stronger emotional or interpersonal components experienced the greatest employment growth. These are distinct outcomes: the tasks with the largest employment gains were not necessarily those with the largest wage gains.

The analysis also finds that capital deepening—the increased use of capital relative to labor—occurred in tasks with low social perceptiveness. In the study’s framework, tasks involving greater social perceptiveness are harder to replace with physical capital. This does not mean that interpersonal work is immune to technological change; it indicates that the relationship between technology and labor substitution differs across task characteristics.

Education, meanwhile, showed relatively little variation over time across the labor groups considered, and educational differences among those groups narrowed. The findings suggest that formal education alone may not capture how computerization changes employment and wages. Other features of work, including autonomy and social interaction, can also help explain why outcomes diverge.

What this means for automation planning

The study is an analysis of broad U.S. labor-market patterns, not a forecast for a particular plant, industry or technology project. Its data and methods also have limits: census observations are available by decade, and the occupational categories are built using O*NET data. The findings should therefore inform questions about workforce change, not be treated as a direct prediction of the effects of a specific automation investment.

For system integrators and operations leaders, the task-based approach suggests a practical way to structure workforce discussions:

  • Break roles into tasks. Identify which activities are structured and which require workers to set priorities, respond to exceptions or exercise judgment.
  • Include interaction in the assessment. Consider where coordination, communication or awareness of others’ responses is part of the work.
  • Separate employment from wages. Changes in the number of people performing a type of work and changes in its compensation are related but not identical outcomes.
  • Look beyond formal education. Training plans may need to reflect changing task requirements, not just credentials or job titles.

Conclusion

Computerization reshapes work unevenly. The research links higher wage growth with high-autonomy tasks and stronger employment growth with emotional tasks, while finding that capital deepening has occurred in work with lower social perceptiveness. For industrial decision-makers, the practical lesson is to evaluate automation at the task level. Understanding how much discretion and human interaction a task involves can make workforce planning more precise than relying on occupation labels or education alone.

Access Official Technical PDF

About The Author