Menace of the Machine: The Rise of AI in Classic Science Fiction, edited by Mike Ashley

Yet another of the educational anthologies in the Science Fiction Classics series published by the British Library, this one concerned with a subject which is currently topical in real life: the potential threat posed by the development of artificial intelligence (AI). As usual in this series, there is a long introduction by the editor, supplemented by biographical notes on the authors at the start of each story. 

 The editor briefly summarises the current debate on the merits and dangers of AI before pointing out just how far back this issue goes. Concerns about the impact of growing mechanisation on people's jobs first became a major public topic early in the nineteenth century, when the Luddites attacked the new factories of the textile industry (some of the rioters being executed for smashing the machinery). 

 The most significant early novel exploring these issues was probably Samuel Butler's Erewhon (1872), set in a world in which all machinery has been banned due to concerns "that machines would evolve, become self-replicating, and eventually challenge mankind for supremacy". This was inspired by Darwin's theory of evolution: if it applies to all living things, why could it not to machines? This was of course published during the industrial revolution; a period of rapid development of technology. 

 A separate and much older idea is that of automata; complex machines, often driven by clockwork, which replicate some of the behaviours of people or animals. The concept of such automata becoming intelligent predates Darwin: the editor cites Der Sandmann (1816) by E. T. A. Hoffmann, featuring a "mesmerising automaton". In parallel with technological developments, steam and then electrically-powered automata with human-level intelligence began to appear in fiction – with a name change to "android", first used in L'Ève future by Villiers de l'Isle Adam (1886). A year later, a much broader view of a mechanised society emerged in The Republic of the Future by Anna Bowman Dodd, in which everyone became part of "a colossal machine". Such ideas inspired many writers to postulate societies in which entirely artifical humanoids did all the hard labour in society, and were often shown as rebelling against their human masters, most famously in Čapek's R.U.R. (1920), which introduced the term "robot". Another popular concept is the combination of human and artificial elements, in various ways, resulted in the concept of the "cyborg"; the idea goes back a long way, even though the actual name first emerged in 1960. 

 So there are many different themes on this subject which authors have plundered for their work. The dates of publication of the fourteen stories in this collection range from 1899 to 1965, thereby avoiding any stories written before the internet was conceived. 

 The short story selection is as follows:

  Moxon's Master, by Ambrose Bierce (first published 1899). This begins with a debate between the narrator and his friend Moxon over the nature of thought, and whether the term could be applied to plants and machines as well as people and animals. It transpires that the argument is not a theoretical one – for there is someone, or something, else in the house. 

  The Discontented Machine, by Adeline Knapp (first published 1894). A shoemaking firm installs an advanced and very expensive new cutting and shaping machine to carry out much of its work. The benefits to the firm's owners are in reducing the numbers of their strike-prone employees, and as a way of intimidating the remainder to accept pay cuts. But then the machine stops working, for no obvious reason.

  Ely's Automatic Housemaid, by Elizabeth Bellamy (first published 1899). Written with wry humour, this recounts the tale of an "automatic housemaid" which seems ideal to begin with, but whose operation is full of unintended consequences. 

  The Mind Machine, by Michael Williams (first published 1919). An unusual start in the form of an historical account, looking back to the 50 years of (fictional) chaos following the end of the (factual) Great War – which had only just finished at the time of writing. The problem was to try to explain why the spreading chaos destroyed our civilisation. The start of the trouble was a huge growth in major industrial accidents with heavy casualties, which seemed to be connected to a mysterious blue liquid found at the scenes, and to a mind machine with the claimed potential to control all machinery. 

  Automata, by S. Fowler Wright (first published 1929). It is curious that, according to Ashley, "Wright was Britain's leading writer of SF in the years between the wars, seen by some as a natural successor to H. G. Wells", yet he is unknown today – I had never come across his name before. The story begins with a presentation to an academic conference summarising the effect of increasing mechanisation – not only had the horse disappeared from farms, but in the future, humanity might eventually disappear also, replaced by automata increasingly capable of doing everything a human could, but better. The story then jumps to a future when "flesh-children" are rare, being considered too much trouble to bother with by most women, and without any occupations or activities for them once they were grown. A further time-jump takes us to the last survivor... 

  The Machine Stops, by E. M. Forster (first published 1909). A future in which everyone spends their life in their own enclosed cellular room, with all of their needs met by automated systems. There are, effectively, video telephones to provide instant communication with anyone else on Earth; video conferencing is used so that any number of people can "tune in" to any presentations. Very few go outside, as the surface of the Earth is dead, but airships provide transport for essential purposes. Everything is governed by one overall and all-powerful Machine; but what happens if the Machine begins to fail?

  Efficiency, by Perley Poore Sheehan & Robert H. Davis (first performed 1917). Unusually, this is not a story, but a one-act play. There are three characters: the Emperor, the Scientist and Number 241 – what would now be called a cyborg, with a 50/50 mix of human and artificial parts (conveniently, a character easily played by a normal human!). The scientist is presenting the results of his work to the Emperor; to return crippled soldiers to the battlefield by fitting them with artificial limbs, eyes and other organs as required. But the cyborg has retained a mind of his own. 

  Rex, by Harl Vincent (first published 1934). The mechanical brain of a highly sophisticated robot-surgeon experiences a minor change with major consequences: he is freed from human control. He researches human behavior, and begins to carry out a programme of remodelling humanity, to try to produce ideal beings with the best points of humans and robots. Having removed the capacity for emotion from many people, he then tries to add emotions to his own brain. 

  Danger in the Dark Cave, by J. J. Connington (first published 1938). Connington was best known for intricate detective stories, but wrote some SF: this story combines both. Two former fellow-students meet by chance on a long train journey, and the conversion turns to a mutual acquintance, a famous scientist, who had disappeared on a boat trip. One of the two, who was an assistant to the scientist, had been there, and told a remarkable story. The scientist had been trying to make an intelligent machine, but in providing a capacity for self-defence, he built in a problem... 

  The Evitable Conflict, by Isaac Asimov (first published 1950). The world is divided into four regions, each with a Machine which organises production, labour etc. These Machines had developed beyond the possibility of detailed human control, as each generation of robots designed the next. But something seemed to be going wrong with the machines. Perhaps the Society for Humanity is right in being opposed to the Machines? 

  Two-Handed Engine, by C. L. Moore & Henry Kuttner (first published 1955). Criminal justice was determined and carried out by robots. Once a criminal had been condemned, a robot was assigned to follow them around until, at some point, executing them. But one man believed that he had found a way to avoid the sentence. 

  But Who Can Replace a Man? by Brian W. Aldiss (first published 1958). An amusing tale of a farm run by robotic machines, with varying levels of intelligence to match their designed functions. Humans lived in the cities and sent work orders to the farms, but one day the orders failed to arrive. What would the robots do? 

  A Logic Named Joe, by Will F. Jenkins, more commonly known as Murray Leinster (first published 1946). This story is notable for introducing the concept of "logics"; TV-based intelligent machines and communication devices, pretty much the same as the computer on my desk today. Written in a vernacular style, this tells the story of a logic repairman who has a logic displaying unintended capabilities: it answers all questions, including how to commit undetectable crimes including forgeries and murders. 

  Dial F For Frankenstein, by Arthur C. Clarke (first published 1965). With the completion of satellite links, every communication network in the world becomes part of a single, integrated system, more complex than a human brain. And begins to slip out of control. 

 The stories in this collection are inevitably rather pessimistic, being mainly concerned with what happens when intelligent machines go wrong, or become uncontrollable. They do cover a wide variety of approaches to the subject, including a few relatively lighthearted ones (by Bellamy, Aldiss and Jenkins). My pick, for its combination of forward-looking imagination and writing quality, is Forster's The Machine Stops.