Before the AI craze, the rise of automatons in the 18th century

Artificial intelligence has now changed the ways in which we define machines, but the idea of a self-operating machine-what is generally called an automaton-is decidedly not new. The term automaton comes from the Greek word automatos, meaning “acting of one’s own will,” and the concept reflects a long-standing interest in machines with life-like qualities. Well before the AI revolution, inventors in ancient Greece and beyond built machines that could run on their own, fascinating and intriguing many of philosophical interest. These antique automated devices thus served as precursors of modern robots, therefore stimulating both wonder and philosophical debate on what life and automatization actually are.
In ancient Greece, there had been frequent uses of automata both in mythology and practical inventions. By the 5th century BCE, Archytas of Tarentum, a mathematician, was said to have constructed a mechanical pigeon powered by steam that could fly for a short period of time. It was presumably one of the earliest recorded flying self-propelled devices and therefore demonstrated some primitive form of robotics.
The Greeks also built complex machines for various applications in temples and entertainments. Hero of Alexandria was a prolific inventor of the 1st century CE who recorded a number of automata in his work “Pneumatica.” His other inventions included a coin-operated machine, a wind-powered organ, and an automaton theatre that enacted entire scenes with mechanical figures. These were precursors of the steam, water, and air engines, testifying to the outstanding knowledge of pneumatics and mechanical engineering possessed by the Greeks.
Automata survived through the ages, with significant developments during the Islamic Golden Age: the 8th to 13th centuries. Engineers such as Al-Jazari built complex machines, including water clocks, fountains, and even programmable humanoid musicians. Their ingenuity married artistry with functionality and inspired European inventors in later centuries. 
By the time of the Renaissance, European engineers were in full drive, tinkering at the boundaries of mechanical ingenuity. This was famously envisioned as a robotic knight by Leonardo da Vinci in the late 15th century, and later, inventors like Jacques de Vaucanson built realistic automata to amaze the public. One of the most famous examples of early 18th-century automata is the “Digesting Duck” by Vaucanson, an automaton capable of flapping its wings, eating, and then giving the impression of digesting food. Although it was, as it later proved to be, an illusion, it engrossed the audience and underlined the period’s preoccupation with imitating life through machines.
During the 18th and 19th century, automata reached new levels of intricacy and artistry. Swiss watchmakers Pierre Jaquet-Droz and Henri Maillardet built mechanical wonders that could draw, write, and play musical instruments. Maillardet’s automaton, as generally known in its common name, “The Writing Boy,” was one of the most famous examples of 19th-century automata. This was created, circa 1800, by the Swiss-born mechanical magician Henri Maillardet, who was active in London. It is a machine of extraordinary artistry and precision, able to both write poems and draw pictures, considered to be one of the most complexly “remembering” of any surviving automaton of its era.
Operating via a complex series of cams, rods, and levers, brass cams sit at the heart of it. The cams were cut out to regulate the direction and pressure of the automaton’s pen in such a way that reproduces handwriting or drawings with precision. In other words, the cams store the “memory” of the drawings and writings, much like a computer would do the same.
This machine, accordingly, is powered by two clockwork motors. One of these operates the movements of the cams, while the other drives the movement-an arm assembly carrying the pen-which automatically lifts it where need be and pushes it down. This is an interplay of three fingers in steel that trace the outlines of cams, making its movements. By so doing, these fingers convert the rotary motions of the cams into the up and down, side to side, and forward and backward motions that make drawing and writing possible​.
This automaton created four different drawings and could write three poems-two in French and one in English. Such versatility and detail had never been seen, setting this automaton apart from all the other mechanical devices of its time. That this could be executed this way testifies to ingenuity and mechanical precision at the time of creation.
Eventually, the writing boy automaton found its way across the Atlantic Ocean to the United States and into the collection at Philadelphia’s Franklin Institute. That institute acquired the machine in 1928, but by then it was disassembled and of unknown origins. It was during this restoration that the researchers found its identity when the automaton wrote at the bottom of one drawing, “Ecrit par l’Automate de Maillardet”. This signature authenticated the machine and confirmed Maillardet to be its creator​.
It was many decades before this automaton, at the Franklin Institute, was restored and cared for by this team, yet it continued to this day to perform its routines. Nowadays, it is one of the most popular exhibits by far, as visitors come to see this marvelous machine in action. It was a repair process which showed not only the technical skill in which this thing was made but also the continuing power of such mechanical marvels.
First demonstrated to the Franklin Institute in Philadelphia, the animatronic boy in Maillardet’s automaton is more than just a curiosity from ages past; it is the technological aspiration for the age in which it was created. Such automatons were devised to copy specific human actions as precisely as possible, the borderland between human inspiration and its mechanical reproduction. In an era before electronic computers existed, this sort of gadgetry represented the state-of-the-art for mechanical computation, encoding complicated actions in a purely mechanical form​.
Nowadays, Maillardet’s automaton is increasingly regarded as one of the first programmable machines. The use of a set of cams for storing a set of instructions can be regarded as one of the very early uses of data encoding, a principle underlying today’s computing. As such, the writing boy serves not only to capture the imagination of those interested in the history of technology but also as a tangible link between the automata of the past and today’s AI-driven robots.
The Maillardet’s writing boy automaton has even found its place in popular culture. The automaton was an inspiration for the movie Hugo, 2011, by Martin Scorsese, who told the story of a young orphan discovering a mechanical man who could write and draw.
Another famous 18th-century automaton was The Turk-an amazing chess-playing machine invented by Wolfgang von Kempelen in 1770. Unlike Maillardet’s automaton, a real mechanical wonder of its time, The Turk was an elaborate hoax. The machine took the form of a life-sized figure in Ottoman costume seated at a cabinet containing a chessboard and was promoted as a self-operating chess player capable of defeating human opponents, reportedly including such well-known personalities as Napoleon Bonaparte and Benjamin Franklin.
But in reality, the operator of The Turk had a human chess master inside its case while it controlled the mechanical arm. Debunked as a hoax, the Turk had nevertheless enthralled audiences throughout Europe and America for over a century and blurred the distinction between human and machine intelligence.
The legacy of automata extends far beyond their mechanical bodies. These early machines not only amazed their audiences but also constituted the very foundation upon which the edifice of modern robotics and artificial intelligence has been built. The principle of programming and automation they had played out underpins even today’s sophisticated technologies.
Automata such as Maillardet’s writing boy and The Turk demonstrated how machines could replicate some of the most complicated and high-order human pursuits, from writing poetry to playing chess. This interest in emulating human skills followed right through to the 20th century, where mathematicians like Alan Turing codified thoughts on artificial computation. Turing’s theoretical model of computation-the so-called Turing machine-made use of ideas on programmed actions and encoded memory that were conceptually similar to the cam-based mechanisms of Maillardet’s automaton​.
These principles pioneered by automata have evolved into sophisticated AI algorithms in the digital era for everything from smartphones to self-driving cars. Unlike their mechanical forebears, today’s AI systems process information at staggering speeds, even learning from their interactions, but still harnessed to the basic idea of programmed actions: executing complex instructions to achieve specific outcomes. Execution of the tasks in a deterministic fashion draws a direct line between the cams and gears of the 18th-century automatons to the code driving modern AI.