Technology, jobs and daily life

Automation and work, how daily life changed, and how automatable a job is.

F12.8Technology and societyGCSE15 min

The largest effects of computing are the ones people stop noticing: how we work, shop, talk and spend our evenings. Exam questions call these cultural impacts. This lesson looks at the change to work, and at what always-connected life does to people, and it measures how automatable a job really is.

Work

Automation has been changing work since the first factory machines, and software has sped it up.

• Jobs removed: routine, repetitive work is automated first, in factories, warehouses, call centres and back offices.
• Jobs changed: most jobs keep their purpose but change their tools. A mechanic now reads diagnostics; a farmer reads satellite maps.
• Jobs created: people are needed to build, run, repair and check the systems, in work that often did not exist a generation ago.

The gain and the loss do not fall on the same people. Someone whose job is automated at fifty does not automatically become a data engineer. Retraining, and who pays for it, is the real argument.

Where work happens changed too: remote and hybrid work saves commuting and opens jobs to people far away or with caring duties, but it can blur work and home and leave people isolated. Gig work brought flexibility and insecurity together.

Daily life

Notice that almost every row has both. An exam answer that gives only one side is half an answer.

How automatable is a job?

Not whole jobs are automated, but tasks within them. Add up the tasks, and you see which jobs change most:

# the two lines every program starts with: the commands, then the robot
from bugbot import *
connect()

# each task: name, share of the job, how automatable (0 to 1)
tasks = [("stacking shelves", 0.4, 0.9), ("helping customers", 0.4, 0.2), ("counting stock", 0.2, 1.0)]
score = sum(share * auto for _, share, auto in tasks)
print(f"automatable: {score * 100:.0f}% of the job")
print("the part that is hardest to automate:", min(tasks, key=lambda t: t[2])[0])

Run this in the simulator

The tasks left over are usually the human ones: judgement, care, persuasion, and dealing with the unexpected.

Task: how automatable?

For each job in jobs, work out its automatable share: the total of each task's share times how automatable it is. Print <job>: <n>% rounded to the nearest whole number, sorted from the most automatable to the least. Add AT RISK to any job of 60% or more. At the end print most human task: <task>, the task with the lowest automatable value across every job.

# the two lines every program starts with: the commands, then the robot
from bugbot import *
connect()

# job: [(task, share of the job, how automatable 0 to 1), ...]
jobs = {
    "shop assistant": [("stacking shelves", 0.4, 0.9), ("helping customers", 0.4, 0.2), ("counting stock", 0.2, 1.0)],
    "nurse": [("taking notes", 0.3, 0.7), ("caring for patients", 0.6, 0.05), ("ordering supplies", 0.1, 0.9)],
    "warehouse picker": [("finding items", 0.6, 0.95), ("packing", 0.3, 0.8), ("checking damage", 0.1, 0.4)],
}

Challenges

1. Which job changes least, and what does its safest task have in common with the others' safest tasks?
2. Add a task of your own to the nurse's job. Does the verdict change?
3. Name a job created by computing in the last twenty years, and one it removed.