The Workers Training Their Replacements: How AI Could Reshape Skilled Labour

Artificial intelligence is no longer confined to chatbots, software and digital workplaces. It is increasingly entering the physical world through cameras, sensors, robotics and computer-vision systems that can observe and analyse how people perform their jobs. In factories, warehouses and other industrial settings, machines are learning to recognise movements, identify patterns and replicate specific tasks.

For workers, this brings both opportunities and concerns. Automation can make workplaces safer, reduce repetitive physical labour and improve productivity. At the same time, it raises an unsettling question about the future of skilled employment: if machines can observe experienced workers, collect data about their actions and learn to perform the same tasks, could those workers eventually be replaced by the systems they helped train?

This question matters especially for skilled workers whose expertise has been built through years of hands-on experience. Much of their knowledge is practical rather than written down, making it difficult to reproduce through conventional training. Artificial intelligence and robotics could change that by converting human experience into usable data.

The Problem: When Human Skill Becomes Machine Data

Skilled workers often develop abilities that cannot easily be explained in a manual. A textile worker may know how to handle different fabrics without consciously considering every movement. A machine operator may recognise a problem from a particular sound or vibration. An experienced technician may know where to look for a fault because of years spent working with similar equipment.

For a long time, this practical knowledge stayed with the worker. Companies could train new employees, but it was difficult to replicate the experience of someone who had spent decades doing a particular job. That made skilled workers valuable and gave their expertise some protection from automation.

AI is beginning to change this relationship. Cameras can record workers performing tasks, while sensors capture information about movement, pressure, speed and position. Software can track errors, successful outcomes and the time needed to complete a process. AI systems can then analyse vast amounts of this information to identify patterns in how the work is done.

As a result, knowledge that once existed mainly in a worker’s experience can gradually become a body of digital data. Once captured and analysed, that data can be used to automate tasks, train systems and optimise production.

Which Skilled Workers Could Be Most Affected?

Manufacturing and textile workers are among those most likely to experience this transformation. Garment workers, for instance, may perform repetitive cutting, stitching, sorting or quality-control tasks. Manufacturing employees may assemble components, operate machinery, inspect products or package finished goods. These processes often take place in controlled environments, where movements can be observed and repeated.

Warehouses are another important example. Workers may spend their days picking products, sorting packages, moving materials and preparing orders. Such activities are increasingly supported by computer-vision systems, automated sorting equipment and robots. As these technologies improve, some tasks now performed by people may require fewer human workers.

The food-processing industry could face similar changes. Sorting, packaging and inspection are particularly suited to systems that combine cameras, AI and robotics. Construction may also be affected as robots become more capable of moving materials, drilling, laying bricks and performing other repetitive physical tasks.

Skilled technicians present a more complex case. Their work often involves diagnosis and judgment, not just repetition. Yet if their repair procedures, decisions and results are systematically recorded, AI systems could learn to perform parts of those processes, too. This does not mean an entire profession can be automated overnight. More often, it is individual tasks within a profession that are gradually handed over to machines.

How AI and Robotics Could Learn from Workers

Physical AI depends heavily on data. A robot does not simply need to know what a finished product looks like; it may also need to understand how an action is performed and how to respond when conditions change.

That is where workplace data becomes important. A camera can record a worker’s hand movements. Sensors can measure the force applied during a task. Other systems can track the position of objects, the time taken to complete an activity and whether the final result meets the required standard. When thousands or even millions of observations are collected, AI systems can begin to identify patterns.

The process resembles a form of digital apprenticeship. Instead of a new worker learning directly from an experienced employee, an AI system can analyse countless examples of how skilled people perform a task. Engineers can then use that information to build automated systems capable of handling some part of the same process.
A machine does not necessarily need to reproduce every decision a human makes. It only needs to perform the parts of a job that can be measured and repeated reliably. That is why repetitive, predictable tasks are likely to be automated sooner than work requiring constant judgment and adaptation.

The Impact on Workers

One of the most significant consequences could be a reduction in the number of people needed to perform certain types of work. Automation does not always make an entire occupation disappear. More often, companies automate individual tasks and reorganise the workforce around the responsibilities that remain.
For example, a factory that once employed a large team to assemble and inspect products might eventually use machines for assembly and automated vision systems for inspection. Human employees may still be needed to monitor production, address unexpected problems and maintain equipment—but fewer people may be required to carry out the original manual tasks.
This shift could improve productivity and make some workplaces safer. Machines can take over tasks involving dangerous equipment, repetitive movements or physically demanding conditions. Companies may also benefit from more consistent production and fewer errors.
But the transition could be difficult for workers whose livelihoods depend on manual skills. Someone with decades of manufacturing experience cannot automatically move into an AI-related role simply because their previous work has been automated. New technical jobs may emerge, but they often require different qualifications and education and that creates a gap between the work that disappears and the work that replaces it.

The Need for Reskilling and Worker Protection

The answer to automation cannot simply be to halt technological development. AI and robotics are likely to become more common in manufacturing, logistics, construction and other industries because they can offer significant economic and safety benefits.
A more practical response is to prepare workers for changing roles. Companies introducing automation could invest in training programmes that help current employees operate, supervise and maintain automated systems. A manufacturing worker, for example, might move from assembling products by hand to operating robotic equipment or monitoring an automated production line.
Governments and educational institutions also have an important role to play. Vocational programmes may need to combine traditional trade skills with technological training. Courses in robotics maintenance, industrial automation, computer-controlled machinery, data handling and AI-assisted production could become increasingly valuable.
The goal should not be to turn every skilled worker into an AI specialist. It should be to give people a fair opportunity to develop the technical knowledge they need to remain valued in workplaces where humans and machines increasingly work side by side.
Companies must also consider how they collect and use data generated by workers. When employees are recorded and their movements analysed, they should be told what information is being collected, why it is needed and how it might be used. Human expertise should not become an invisible resource—captured and converted into technology without regard for the people who contributed it.

The Future of Skilled Labour

It would be premature to conclude that AI will eliminate skilled workers altogether. Physical environments are often unpredictable, and machines still struggle with many situations that people handle naturally. Human workers can adapt to unusual circumstances, communicate with others, make judgments and respond to problems that were never included in a system’s training data.
The more likely future is a gradual transformation of skilled work. Some repetitive tasks may disappear, while new responsibilities emerge in supervising, maintaining and improving automated systems. The definition of a skilled worker may change rather than vanish.
The key question is whether workers will receive enough time and support to make that transition. If companies and governments treat automation solely as a way to reduce labour costs, technological progress could lead to widespread displacement. If it is accompanied by serious investment in education and reskilling, automation could instead remove hazardous or monotonous tasks and help workers move into higher-value roles.

Conclusion

The future of work is unlikely to arrive in a single moment when machines suddenly replace people. More likely, it will unfold gradually, as individual tasks are automated one by one. Cameras will record workplace processes, sensors will capture physical actions, and AI systems will grow increasingly capable of analysing that information.

Perhaps the most profound change is that human expertise itself is becoming easier to capture. Knowledge that once existed only in the hands and minds of skilled workers can now be converted into digital data and used to train machines.

That brings both opportunity and warning. Technology can make work safer, faster and more productive but it can also reduce demand for the very skills that helped create those gains.
The challenge of the coming decades is not simply to decide whether machines should replace people. It is to ensure that, when machines learn from human workers, the people who taught them are not left behind.