Can artificial intelligence and automation reduce the human factor and creativity in the industry?

Can AI and automation reduce the human element and creativity in industry?

Automation that excludes human judgement deserves scrutiny

Hello everyone. Poorly designed AI and automation can narrow workers' scope to apply knowledge, ask questions and develop solutions. Reduced creativity is not an inevitable result of every installation, however. The central question is which decisions are taken away and what remains for employees. When Industry 4.0 marketing promises more creative work, it should explain the time, authority and training provided. Removing repetitive tasks does not automatically create conditions for creativity.

Editor's note: A human approval button does not establish meaningful human oversight. If employees lack the knowledge, time or authority to reject a recommendation, approval may be merely formal. Leaving people responsible while removing their ability to decide is a serious design problem.

How can skills erode?

When employees spend years following system instructions, opportunities to use their understanding of cause and effect may diminish. Yet the same people may suddenly be expected to act as experts during an unfamiliar failure. This contradiction matters in discussions of skills and job security. EU-OSHA's robotics review addresses preventing skills loss and involving workers during implementation.

The answer is not unlimited permission to change settings. Provide explanations appropriate to each role, practical training and safe opportunities to learn. Standard work can support consistent quality, while a route for questioning and improving the standard should remain available when problems arise.

Experienced workers may know what the database does not

Imagine a hypothetical line with intermittent part-detection problems. Software classifies them as feeding issues, while an experienced technician notices they occur only after washdown. Looking solely at a SCADA chart without examining the sensor's position, cable and surrounding conditions may leave the diagnosis incomplete. This example illustrates context, rather than diagnosing a particular fault.

Likewise, an oscillating temperature may call for checking measurement and process conditions before changing PID settings. Unwritten experience can be shared through fault notes and joint investigations. Data analytics can support that transfer; it cannot be assumed to know information that has never been recorded.

Creativity means more than generating more ideas

Industrial creativity can mean a fixture that is easier to change, a clearer alarm or a layout preventing recurring faults. If everyone accepts the first suggestion from the same system, exploration may narrow. AI can also help people think by offering alternatives. Saying that a machine can never generate a new idea is not a sound technical explanation either.

In Doshi and Hauser's 2024 short-story experiment, stories written with AI-generated ideas received higher creativity ratings but were more similar to one another. This does not establish an equivalent reduction in creativity on factory floors. However, improving individual outputs and increasing the diversity of options are different measures. Engineering alternatives also need assessment against manufacturing, maintenance, safety and cost constraints.

Protect teamwork and room for judgement

Turning every problem into an individual performance score may discourage people from reporting mistakes or suggesting improvements. The OECD's job quality review discusses work intensity and autonomy risks arising from poor implementation. Modern screens alone do not deliver job satisfaction. Operators, maintenance staff and engineers need time to investigate shared problems together.

Be clear about the use of employee data, and distinguish machine downtime from individual performance. Where staffing cuts and fear of job loss exist, telling people to be more creative is insufficient. Plan training, new responsibilities and ownership of the improvement process explicitly.

When can AI expand people's scope?

Reducing time spent preparing routine reports, searching records or comparing drafts can free people to solve problems. That space will not emerge if every saved minute becomes a tighter production target. After a claim of time savings, examine how the time is actually used.

A digital twin or simulation may help examine an idea before trying it on the physical line, subject to the model's limitations. Select proposals for the value they provide to users, rather than their visual appeal. Freedom to experiment does not mean bypassing safeguards or making uncontrolled process changes; retain appropriate access and safety boundaries.

Evaluate the human contribution too

Alongside cycle time and scrap, success evaluation should examine recurring faults, implemented improvements, workers' ability to explain problems and their decision authority. Counting more suggestions alone does not demonstrate creativity; follow the outcomes of useful changes. Training and knowledge transfer must not disappear from the project cost.

Competitive advantage should not depend on discarding a business's accumulated experience. A solution that leaves people merely compensating for system weaknesses is incomplete. Assess technology partly by how well it brings employees' knowledge, objections and capacity to improve into production.

Related topics

What the technologies do and where they fall short
Effects on production processes
Steps for implementation
The role of industrial IoT
Implementation challenges for SMEs
Effects on the labour market