The Last Person Who Knows How
By Ayman Fouad Abdelgawad
Automation can save effort while dismantling the capacity to act without it.
Imagine a team that once checked a calculation by hand. Software now produces the answer instantly. Experienced staff initially recognise implausible results. Years later, the calculation is still fast, but those who understood its machinery have left. Using the system has displaced some of the knowledge needed to question it.
This is not a case for performing every task manually. Tools make difficult work possible and release attention for better purposes. Keeping every obsolete skill would consume resources. The question is which capabilities remain essential when the ordinary tool fails or produces a convincing mistake.
A skill's value exceeds its frequency of use. Some knowledge functions like an emergency exit: its importance becomes visible under conditions in which everyday efficiency offers little guidance. A workforce can appear more productive while becoming less able to recognise the boundaries of its instruments.
Retained understanding need not mean reproducing every calculation. People may need to estimate an expected range, recognise contradictory inputs, explain assumptions and know when to stop relying on an output. These are forms of judgement rather than nostalgia for difficult work.
The strongest objection is that human checking can also be unreliable. A ceremonial second look adds confidence without protection. That criticism defeats meaningless supervision, but strengthens the case for explicit, practised competence. Someone must understand what a useful check would detect.
Organisations should ask which abilities are disappearing as systems improve, and whether those losses are acceptable. Exercises under unusual conditions can reveal weaknesses hidden by routine success. A fallback plan nobody has practised is closer to a reassuring document than a capability.
The final expert rarely disappears dramatically. They retire, transfer or stop being asked. By the time their absence matters, an institution may no longer know enough to describe what it has lost.
Editor's reflection
I welcome tools that remove drudgery. My hesitation begins when convenience weakens our ability to distinguish an answer from a plausible display. The skill worth protecting is the ability to recognise when help has stopped helping.