Before the word “robot”
People imagined artificial helpers long before engineers could build them. Myths from several cultures describe mechanical servants or statues that move. Real devices came too: ancient Greek writers described clever mechanisms driven by water, steam or weights, and the Hellenistic engineer Hero of Alexandria is often associated with such designs. In the medieval Islamic world, the engineer al-Jazari described programmable-looking water clocks and mechanical figures in the early 1200s.
From the 1500s to the 1700s, European makers produced automata, intricate clockwork figures that could write, play music or imitate animals. These machines followed fixed sequences built into gears and cams. They could not sense their surroundings or change what they did, so they were entertainment more than robots in the modern sense.
Where the word came from
The word “robot” comes from Czech theatre. Karel Capek’s 1920 play R.U.R. used the term for artificial workers, derived from a Czech word associated with forced labor. In the play, the robots were closer to manufactured people than to metal machines, but the word spread quickly. The word “robotics” was later popularized by the science fiction writer Isaac Asimov in the 1940s, who also wrote his well-known Three Laws of Robotics as story devices. They are fiction, not engineering standards.
First machines that sensed and reacted
In the late 1940s, the neurophysiologist W. Gray Walter built small wheeled “tortoises” that responded to light and obstacles. With very simple electronics they showed behavior that looked purposeful. This was an early demonstration that a machine could react to its environment instead of just replaying motions.
The gap between science fiction and practice was still wide, because real robots needed three things: actuators to move, sensors to perceive and a controller to decide. Postwar developments in electronics, feedback control and remote manipulators for handling radioactive material provided the first two. Controllers had to wait for computers.
Industrial robots
The first big commercial success came in factories, where tasks are repetitive and the environment can be controlled. George Devol patented a programmable transfer device in the 1950s, and with Joseph Engelberger he helped develop the Unimate. A Unimate arm was installed at a General Motors plant in New Jersey in 1961 to handle hot metal parts, a task that was dangerous for people.
During the 1970s and 1980s, robot arms spread through car manufacturing, particularly for welding and painting, and Japan became a major adopter. Standardized arms with several jointed axes, along with programming interfaces that let workers “teach” a motion, helped companies deploy them. These machines usually work inside safety cages, doing the same motion with high repeatability. Today’s smart factories build on that foundation by adding sensors and data analysis.
Robots that move and plan
Research labs asked a harder question: could a robot work in a place it had never seen? In the late 1960s, the Shakey project at the Stanford Research Institute produced a wheeled robot that used a camera, bump sensors and planning software to move around rooms and push objects. It was slow, but the project contributed methods in planning and path search that are still taught.
Researchers also learned that tasks that seem easy to people, such as walking over uneven ground or picking up a cup, are hard for machines, while tasks like arithmetic are easy. This observation is sometimes called Moravec’s paradox. In the 1980s, some roboticists argued for simple reactive behaviors layered together instead of detailed world models, an approach that influenced later designs.
Robots leave the factory
By the 1990s and 2000s, cheaper processors, better batteries and improved sensors let robots go into new settings. Surgical systems let doctors control instruments with fine precision. Warehouse robots moved shelves and packages. Robotic vacuum cleaners reached homes in the early 2000s, showing that a modest, well-defined machine could find a market. Much of this became possible with advances in semiconductor chips and compact lithium-ion batteries.
Robots also went where people could not easily follow. Rovers have explored Mars, and space robotics now supports tasks on orbiting spacecraft and stations. Underwater vehicles and bomb-disposal machines, usually remotely operated, extend human reach into dangerous places.
Learning and seeing
In recent years, progress in machine learning and computer vision has helped robots recognize objects and handle varied situations more flexibly. Self-driving cars can be viewed as robots as well, combining many sensors with planning software. Progress has been real but uneven, and many demonstrations still work best in controlled conditions. Claims about what robots will do in the near future deserve some caution.
Recurring themes
Several patterns repeat across robotics history. Robots succeeded first where tasks were repetitive and dangerous or dull. The hardest problems turned out to be perception and handling the unexpected, not raw strength. Advances usually followed improvements in other fields: electronics, control theory, computing and materials. And public imagination, shaped by stories, has often run ahead of what machines can actually do. Understanding that gap is one of the best ways to read any new announcement about robots, including those that affect jobs through workplace automation.