Why be people led?

The suggestion that Industry 4.0 should be people-led invites the question of why this might matter and how it might contrast with prior practice. A quick look back can begin to frame answers to these questions.

Learning from the past

The Industrial Revolution provides the prime example of machine-centric implementation. The nature of factory work revolved around maintaining and supplying the machines with resources like oil, coal, and raw materials were continuously fed into the machines. As a visit to one of our preserved mills makes clear, the machine’s needs, rhythms, and peculiarities dictated not just the work of humans but nearly everything about their pattern of life.

By the time of mass production (Industry 2.0), the focus shifted towards efficiency and productivity. Under Taylorism, there was a growing interest in how machine operators could align with the efficiency and optimisation of mass production systems through carefully designed work processes. The Gilbreths’ work* during this era is emblematic of this shift. They filmed human workers to conduct innovative time and motion studies, leading to minute optimisations of their work. The human worker, subjected to the same kinds of analysis developed for understanding production line efficiencies, became subject to the same kinds of design essentially becoming part of the machinery. Arms and legs, backs and hands are essentially being thought of as if they were themselves just another set of levers, motors and actuators to be organised.

There is some evidence that in “Industry 3.0” (the era of electronics, computers, and the internet) this tendency to analyse humans as if they were the technology of the day has continued. Emerging forms of online work, where, for example, tasks are broken down into manageable pieces and allocated by algorithms are sometimes couched in the language of “human processing units” producing “human computation”. The Amazon crowdsourcing platform “Mechanical Turk” is an example of this trend, named after the famous 18th-century chess-playing automaton (which was secretly operated by a hidden human) that reportedly defeated Napoleon. The metaphor here is that to the user, it appears and behaves like a machine, but in reality, it is a human doing the work. More generally, we begin to see how technologies like email, PowerPoint and word processing begin to shape cognitive activities of writing, communicating and perhaps even thinking. The advent of online ‘platform work’ apps (e.g., fast food delivery work, taxis) featuring small bites of work issued by algorithms again puts us in mind of the ferocious and hungry mills that demanded constant feeding with human effort.

Lessons for the future

From this condensed historical account, we can draw several conclusions for the future of Industry 4.0.

First, if we design around technology (“staff the equipment” rather than “equip the staff”), the human operator begins to mimic, by design, the machine they are operating. This was evident in the industrial age with physical labour being understood as if it were mechanics and in the digital age with cognitive labour starting to be considered as it were the output of a computer.

Second, this phenomenon seems to occur because the analytical frameworks used for designing and understanding machines are then applied to the human operators. This likely arises from the fact that designers generally find it easier to build homogeneous systems composed of like-with-like components.

Third, being people-led is, therefore, partly a challenge in understanding how to design and implement heterogeneous systems of people and machines, where people are allowed to be people. This is sometimes referred to as a “socio-technical system”. People are generally quite good at being human(!) and bring a range of skills like improvisation, imagination, foresight, reflection, instinct, and insight not found in silicon or steel. These skills have considerable value to industry but being unique to people, within a machine-led framework cannot be identified or valued as they are simply out of scope, squandering potentially massive benefits that might otherwise be unlocked.

A final reflection is that while it may have been more feasible historically to make humans act like machines than to make machines act like humans, the advent of AI and deep learning, particularly large language models and vector flow (i.e., ChatGPT and Dall-E), has made this observation less certain. It may be possible to reconcile human/machine heterogeneity with AI implementations designed to help people work better; a people-led approach where machines follow.

* Lillian and Frank Gilbreth, both industrial engineers and a married couple, were known for incorporating their methodologies into their domestic life. Their children wrote the humorous novel “Cheaper By The Dozen” (1945), which has since been adapted into several films, about their parents’ unique and highly efficient approach to managing a household and conducting occasional unexpected experiments. If you’re familiar with the foot-pedal bin, refrigerator door shelves, and the well-known “work triangle” between the stove, fridge, and sink that is often highlighted when selling a new kitchen, then you’re experiencing Lillian Gilbreth’s legacy on a daily basis.

Author’s profile

Dr Robert Houghton, an Associate Professor in the Faculty of Engineering at the University of Nottingham and a member of the Human Factors Research Group (HFRG). His work revolves around understanding the human factors in innovative digital services and technologies.

If you would like to know more about this research, please email p-ld@bath.ac.uk.