Skip to content
LumenaryAcademy
Start 45 min

Robots and people

Working safely beside robots, robots in care and at home, jobs, privacy, weapons, and who is responsible when a robot gets it wrong.

Code tested · 2AI-draftedAwaiting review by a roboticist

Your guides: Chloe · The Builder, Gideon · The Safety Engineer, Margaret · The Experimenter

MotivationH

Sharing the world with robots

For most of their history, robots worked in factories, behind fences, where very few people ever came near them. That is changing. Robots now vacuum living rooms, deliver medicines along hospital corridors, carry parcels on pavements, milk cows, help surgeons and stand beside people on production lines. As robots come out of their cages, new questions come with them. How do we keep people safe around machines that are strong and fast? Should robots help care for older people? What happens to people's jobs? Who sees what a home robot sees? Should machines ever be allowed to choose targets in war? And when a robot causes harm, who is to blame?

These questions have no single answer to look up. This lesson gives you the facts to think about them, and ideas to discuss with friends, family and classmates. The robot engineers of the future, perhaps including you, will have to answer them in their own work.

ExplanationH

Working safely beside robots

Why factory robots are fenced

A large industrial robot arm can swing a car door through the air faster than you can run, and it can push with enough force to crush a person. Just as important, a traditional factory robot does not know people exist: it follows its program exactly, and if someone is standing where the program says to move, it moves anyway. That is why such robots work inside a cell: a fenced area with gates. The gates are interlocked, which means that opening one automatically stops the robot. When engineers must work inside the cell, for example to teach the robot a new movement, it runs at a slow, reduced speed, and they hold a handheld controller with a switch that stops it the instant they let go, or grip too hard in panic.

International standards, above all ISO 10218 for industrial robots, set out how robot cells must be designed and checked, and in many countries the law requires them to be followed. The first step of any robot installation is a risk assessment: listing everything that could go wrong, how likely and how serious each danger is, and how it will be prevented.

Definition— emergency stop

An emergency stop is a big red button, usually on a yellow background, that any person can press to stop a machine at once. Pressing it brings the machine's motors to a halt and cuts their power, through circuits designed to keep working even if the computer fails, and the button stays pressed in, so the machine cannot start again until someone deliberately releases it and resets the system.

Collaborative robots

Since the late 2000s, companies in many countries, starting with Universal Robots in Denmark in 2008, have sold collaborative robots, or cobots: arms designed to work right beside people without a fence. They stay safe in two main ways.

  • Power and force limiting. A cobot is light, rounded and moves gently. It constantly measures the forces in its joints, and if it bumps into something unexpected, such as your arm, it stops or backs off. Its speed and force are kept low enough that a bump does not cause injury.
  • Speed and separation monitoring. Sensors such as laser scanners or cameras watch the area around the robot. When a person comes closer, the robot slows down; when they come very close, it stops; when they leave, it speeds up again.
robot and scannerstop zone: 1 mslow zone: 2 ma person walking inat up to about 1.6 m/s
Speed and separation monitoring, seen from above. Inside the outer ring the robot slows down; inside the inner ring it stops. The rings must be big enough for the robot to slow or stop before a walking person can reach it.

Here is the logic of speed and separation monitoring as a short Python program. The numbers are examples; a real cell gets its numbers from a careful safety assessment. Press Run.

python
# Speed and separation: a robot slows down, then stops, as a person comes near.
# The zone sizes and speeds are example numbers; a real robot cell gets
# its numbers from a careful safety assessment.

def allowed_speed(distance):
    # distance: metres between the person and the robot
    if distance > 2.0:
        return 1.0          # nobody near: full speed (metres per second)
    elif distance > 1.0:
        return 0.25         # slow zone: a quarter speed
    else:
        return 0.0          # stop zone: the robot stands still

# A worker walks up to the robot, waits while it is stopped, then leaves.
distances = [4.0, 3.2, 2.4, 1.8, 1.2, 0.8, 0.8, 0.8, 1.5, 2.5]

for second, distance in enumerate(distances):
    speed = allowed_speed(distance)
    if speed == 1.0:
        zone = "full speed"
    elif speed > 0:
        zone = "SLOW"
    else:
        zone = "STOPPED"
    print(f"t = {second} s   person at {distance:.1f} m   robot speed {speed:.2f} m/s   {zone}")
Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.

The robot runs at full speed while the worker is more than 2 m away (seconds 0 to 2), slows to a quarter speed at 1.8 m and 1.2 m, stands still for the three seconds the worker spends 0.8 m away, and speeds up again as they leave. It is the same sense–think–act loop as in the first lesson, with safety as its whole purpose.

ExplanationH

Robots in care and at home

The robots most people meet first are at home: robot vacuum cleaners, small, slow and light enough that bumping into a person or a pet does no harm, and robot lawnmowers, which stop their blades the moment they are lifted or tipped.

In hospitals and care homes, robots deliver medicines and meals along corridors, disinfect rooms with ultraviolet light when nobody is in them, and help nurses lift patients, a heavy task that injures many care workers' backs. Japan, where more than a quarter of people are over 65, has led a lot of this work. One well-known example is PARO, a soft robot shaped like a baby seal, developed at Japan's National Institute of Advanced Industrial Science and Technology: it responds to touch and voice, and care homes in many countries use it to calm and comfort people living with dementia.

Robots like these raise real questions, and it is worth talking them through:

  • If a robot does the lifting and fetching, carers may have more time to talk and listen. But a care home could also use robots to employ fewer people. Which happens is a decision made by people, not by robots.
  • Is it kind or unkind to give a lonely person a robot pet? It may bring real comfort; it may also become an excuse for nobody to visit.
  • A robot that helps someone live at home for longer can protect their independence. Who decides what it records, and who it tells if something seems wrong?
ExplanationH

Robots and work

Machines have been changing work for as long as there have been machines. Automatic telephone exchanges replaced the people who once connected calls by hand; lift operators disappeared when lifts got buttons; welding robots took over much of the hot, dangerous, repetitive work on car assembly lines. At the same time, new jobs appeared: people design, build, program, install, repair and supervise the machines, and cheaper products created work elsewhere.

What economists say is more careful than either "robots will take all the jobs" or "nothing will change". Robots usually take over particular tasks rather than whole jobs, especially tasks that are repetitive, predictable, dirty or dangerous. The effects are uneven: some workers lose jobs they were good at and find it hard to start again, while others gain. How fast and how far this goes is genuinely debated, and robots that learn, which you will meet at the end of this course, may change the picture. How the benefits are shared, through training, support for people whose work disappears, and decisions about how robots are used, is a choice for societies, businesses and workers, not something technology decides on its own.

ExplanationH

Privacy

To do their jobs, many home robots build a map of your home, and some carry cameras and microphones. A robot vacuum's map shows how big your home is and how the rooms are laid out; its camera may see whoever walks past. That data is often sent to the company's computers, to improve the product or to let you control the robot from your phone.

That can go wrong. In 2022 journalists reported that photos taken by test versions of one company's robot vacuum, including pictures of people inside their homes, had been shared online by workers hired to label images for training the robot's software. The people photographed had agreed to test the robot, but not to that.

Good questions to ask about any robot with sensors in your home: What does it record? Is the data processed on the robot itself, or sent away? Who can see it, and for how long is it kept? Can I switch the camera off, or delete the map? In many places, laws such as the European Union's data protection rules give people rights to ask exactly these questions.

ExplanationH

Robots and war: the debate over autonomous weapons

Militaries use robots too: to defuse bombs, to search buildings, and as remote-controlled aircraft. Most of these are controlled by people. The hardest question is about autonomous weapons: weapons that, once switched on, choose and attack targets by themselves, without a person deciding each attack.

Some systems already work automatically in limited ways, for example defences on ships that shoot down incoming missiles faster than a person could react. Here are the main arguments people make, described as fairly as possible.

  • Concerns. Many people, including the International Committee of the Red Cross (ICRC), argue that decisions to kill must stay with humans; that machines cannot reliably tell a soldier from a civilian, or one who is surrendering; that when a machine makes the choice, it becomes unclear who is responsible; and that such weapons could make wars easier to start and spread quickly.
  • Other views. Some governments argue that existing laws of war already apply to any weapon, so no new treaty is needed; that machines might in some situations make fewer mistakes than tired, frightened people; and that careful design can keep humans in control.

Governments have discussed the question at the United Nations in Geneva since 2014, under a treaty called the Convention on Certain Conventional Weapons, and in the UN General Assembly. The ICRC and the UN Secretary-General have called for new rules that would forbid some kinds of autonomous weapons and strictly limit others; many countries agree, and others do not. As of late 2026 the talks are still going on, so check the news for where they stand now.

ExplanationH

Robot laws in fiction, and why they are hard to program

Writers imagined robots long before engineers built them. Čapek's play R.U.R., which gave us the word, is about robots rebelling against the people who made them. In the 1940s the American writer Isaac Asimov, who is usually credited with first using the word "robotics" in print, invented his famous "Three Laws of Robotics": roughly, that a robot must not harm a human or let a human come to harm, must obey people unless that would cause harm, and must protect itself unless that conflicts with the first two. In Japan, Osamu Tezuka's comic hero Astro Boy, from the 1950s, showed a robot as a kind friend and helped shape how many people there feel about robots.

Asimov's laws sound like a sensible program. Yet most of his stories are about how they go wrong in situations nobody expected, and that is the useful lesson.

Real robot safety doesn't work by giving robots moral rules in words. It works by engineering: limiting speeds and forces, sensing people and stopping, fences and emergency stops, testing, standards, and keeping humans responsible for the decisions that matter. The questions of right and wrong stay with the people who design, sell and use robots.

ExplanationH

Who is responsible when a robot gets it wrong?

A robot is not a person in law: you can't put a robot on trial. When a robot causes harm, responsibility falls on people and organisations: the company that designed and built it, the programmers, the company or person that bought and set it up, the person operating or supervising it, and the regulators who decided the rules. Often it is shared.

A sad real case shows how. In 2018, in Arizona in the United States, a self-driving test car struck and killed Elaine Herzberg as she wheeled her bicycle across a road at night. Investigators at the US National Transportation Safety Board found a chain of failures: the car's software had detected her seconds before the crash but had not worked out what she was or where she was going; the car's own emergency braking had been switched off during testing, leaving braking to a human safety driver; the safety driver, whose job was to watch the road, was looking at a phone; and the company's safety culture and procedures were weak. The safety driver was prosecuted; the company that ran the tests was not charged with a crime. People still disagree about whether that was fair.

Laws are being updated for robots and AI. In the European Union, for example, a new machinery law in 2023 and new product liability rules in 2024 make clear that software, including AI, is part of a product that must be safe; other countries are writing their own rules. You will see how engineers build safety into robots in Embedded control and functional safety, and how learned robots are tested and deployed at the end of the course, in Evaluation, safety and deployment.

Optional Christian reflection · Made for one another

A reflection from the Christian tradition that inspires Lumenary Academy. It is not part of the mathematics and you are welcome to skip it.

Christians believe that every person is made in the image of God, and so has a worth that does not depend on how useful or productive they are. That belief gives a simple test for any robot: does it serve people's dignity, or treat people as parts in a machine? A robot that lifts a patient so that a nurse can sit and listen to them serves dignity; a robot used so that nobody needs to visit at all may not. Jesus spent much of his time with people others avoided, such as the sick, the poor and the lonely, which suggests that caring for one another is something to protect, not to automate away. You do not need to share this belief to study robotics; it is offered as one way to think about what robots are for.

Common misconceptionsH

Common misconceptions

Misconception

"Robots are dangerous because they might decide to attack us." Today's robots have no wishes of their own. Real dangers are much more ordinary: a strong machine moving where a person happens to be, a sensor that misses someone, a program with a bug, a safety system switched off. That is why safety engineering focuses on speeds, forces, sensing and stopping.

Misconception

"A cobot is safe, so no other precautions are needed." A collaborative robot is designed to be safe in the right application, after a risk assessment that includes the tool, the object it holds and the task. A cobot swinging a sharp blade is not safe just because the arm is gentle.

Misconception

"If a robot makes a mistake, it is nobody's fault." Robots are designed, built, set up and used by people and organisations, and responsibility rests with them. Finding out which people is often hard, which is why investigations, standards and laws matter.

Misconception

"Robots will replace all human jobs soon." Robots are good at particular tasks, especially repetitive and predictable ones, and poor at many things people find easy, such as handling unfamiliar objects in a messy room or comforting someone. Work changes, sometimes painfully, but how it changes depends on choices people make.

Exercises

Exercises

Several of these exercises ask what you think, with no single right answer; for those, the solution gives points worth considering. One asks you to write a little Python.

1. Robots and people

Warm-upHuman

What is the main reason traditional industrial robot arms work inside fenced cells?

Choose one

2. Robots and people

Warm-upHuman

Marta presses the emergency stop on a machine, and it stops. A moment later she takes her hand off the button. What should happen?

Choose one

3. Robots and people

CoreHuman

Safety standards often assume a person can walk towards a machine at 1.6 metres per second. A cobot's sensors and brakes need 0.4 seconds in total to notice a person and come to a complete stop. How many centimetres can the person walk in that time? (That is how much further out the stop zone must reach, before adding a safety margin.)

4. Robots and people

CoreComputational

Write zone(distance) for a robot with two safety rings, where distance is how many metres away the nearest person is. It returns:

  • "full" if the person is more than 2 m away;
  • "slow" if they are more than 1 m away, but not more than 2 m;
  • "stop" otherwise.

The starter forgot the slow zone. Be careful at the edges: exactly 2.0 m is "slow", and exactly 1.0 m is "stop".

Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.

5. Robots and people

CoreHuman

Why can't engineers make a robot safe simply by programming it with a rule like "never harm a human"?

Choose one

6. Robots and people

CoreHuman

A delivery robot in a hospital runs over a nurse's foot because a sensor was mounted at the wrong height, and the hospital had switched off a warning beep because it annoyed patients. Select everyone who might share responsibility.

Choose all that apply

7. Robots and people

CoreHuman

A company is designing a home robot with a camera that helps it avoid socks and cables on the floor. Which design choice best protects the privacy of the people who live there?

Choose one

8. Robots and people

CoreHuman

Which statement best matches what the lesson says about robots and work?

Choose one

Further reading and resources