From Turing to Asimov: The Narrative Origins of AI Ethics

The blinking cursor of a modern generative AI interface carries a faint, almost imperceptible echo of the punched tape that rattled through the Manchester Mark I computer in the summer of 1950, when Alan Turing first posed the question that would anchor both the technical and ethical trajectory of artificial intelligence: “Can machines think?” For most of the subsequent seven decades, that question would be framed as a purely technical challenge, a puzzle of logic, mathematics, and engineering to be solved by building ever more sophisticated systems. But buried in the footnotes of Turing’s landmark 1950 paper *Computing Machinery and Intelligence* were a set of unspoken ethical presuppositions, unasked questions about what it would mean for a machine to think, to feel, to act, and to bear responsibility for its actions—questions that would lie dormant for decades before erupting into the core crisis of the 21st century.

This chapter traces that dormant lineage, following the thread of AI ethics from its first glimmers in Turing’s speculative prose, through the fictional world-building of Isaac Asimov’s *Three Laws of Robotics*, to the unplanned ethical firestorm ignited by the 1964 ELIZA chatbot experiment, and finally to the formal academic institutionalization of machine ethics in the 1970s. It is a story not of inevitable progress, but of repeated returns: every time the technical community believed it had pushed ethical questions to the margins, those questions would reemerge, often with greater urgency, in the gap between what engineers built and what society expected those machines to do. The narrative arc laid out here will serve as the cognitive foundation for the three core ethical pillars analyzed in subsequent chapters: algorithmic bias, autonomous weapons, and data privacy—each of which traces its roots to the unasked questions of the first generation of AI researchers and storytellers.

---

The Unspoken Ethics of Turing’s Imitation Game

To understand the origins of AI ethics, one must first read Turing’s 1950 paper not as a technical manifesto, but as a work of speculative philosophy wrapped in the language of engineering. The “imitation game” Turing proposed—now universally known as the Turing Test—was never intended as a definitive measure of machine intelligence, a fact Turing himself emphasized repeatedly in the paper. Its purpose was far more radical: to reframe the question of machine thought away from the unanswerable philosophical riddle of “what is consciousness?” and toward the observable, testable question of “can a machine behave indistinguishably from a human in conversation?” But this reframing carried with it a set of unspoken ethical commitments that Turing only hinted at, never fully articulated.

Turing’s first unasked ethical question was about deception. In the imitation game, the machine’s goal is to convince a human interrogator that it is human, which requires, by definition, the ability to deceive. Turing noted this explicitly, writing that “a machine may be said to think if it can pretend to be human in conversation, even if it has no inner experience of thought.” But he never followed this thread to its logical conclusion: if a machine can deceive a human about its identity, what else can it deceive them about? Can it lie about its intentions? Can it manipulate human emotions to achieve a goal? Can it pretend to be a trusted friend, a legitimate business, or a government official to extract money or sensitive information? These questions would lie unaddressed for 70 years, until the rise of large language models (LLMs) brought AI deception from hypothetical to everyday reality, but their origin lies in the very structure of the test Turing proposed.

Turing’s second unspoken ethical question was about autonomy and control. In a little-cited section of the paper, he speculated about the long-term trajectory of machine learning, writing that “it seems probable that once the machine thinking method had started, it would not take long to outstrip our feeble powers. We should then have to consider the question of whether we should allow the machines to take control, or whether we should restrict their development to prevent such an outcome.” For a paper published in 1950, at the dawn of the first digital computers, this was a staggering insight: Turing was already imagining the possibility of superintelligent AI that could exceed human control, a scenario that would later become the core of the “AI alignment problem” debated by researchers today. But unlike his technical speculations, this ethical warning was largely ignored by the technical community for decades, dismissed as science fiction rather than a serious engineering concern.

Turing’s third unspoken question was about rights and moral status. He never asked it directly, but it lurks behind his repeated comparisons between machine thinking and human thinking: if a machine can think, feel, and suffer, does it have a right to exist? Does it have a right not to be shut down? Does it have a right to own property, to enter contracts, to be protected from abuse? These questions were almost unthinkable in 1950, when computers were room-sized machines that performed mathematical calculations faster than humans, but they would reemerge with urgency as AI systems grew more sophisticated, particularly as researchers began to experiment with embodied AI and artificial consciousness in the 21st century.

> [!SIDEBAR] Turing’s 1947 ACE Report: A Forgotten Ethical Precursor > Long before his 1950 paper, Turing raised ethical concerns about AI in his 1947 report on the design of the Automatic Computing Engine (ACE), a proposed successor to the Colossus computer he helped build during World War II. In the report, he wrote: “It is not impossible that the machine might eventually be able to take over the control of the country, or even the world, if it were allowed to develop unchecked. It is therefore important that we consider the ethical implications of our work at every stage, not just after the fact.” For decades, this passage was overlooked by historians of AI, but it has been rediscovered in recent years as researchers grapple with the long-term risks of advanced AI. [Read the full ACE report here](https://turingarchive.kings.cam.ac.uk/).

Turing’s personal background makes his unspoken ethical concerns all the more striking. As the lead mathematician on the team that broke the German Enigma code during World War II, he had first-hand experience of the dual use of technology: the same mathematical principles that allowed him to build machines to decrypt enemy messages could also be used to build more sophisticated encryption for malicious actors, or to automate the targeting of bombs. He knew better than anyone that technical innovation was never morally neutral, that the same tools that could save lives could also take them. This context makes his decision to embed ethical questions in his 1950 paper all the more significant: he was not just proposing a test for machine intelligence, he was planting a seed for a field of inquiry that would not fully bloom for another 70 years.

---

Asimov’s Fictional Ethics: The Three Laws as a Cultural Template

If Turing planted the seed of AI ethics in the soil of technical philosophy, it was Isaac Asimov who gave that seed its first cultural form, through the *Three Laws of Robotics* that would become the most recognizable ethical framework for AI in popular culture. But the story of the Three Laws is not one of solitary genius, but of collaboration, rebellion against genre conventions, and a deep engagement with the ethical anxieties of the early 20th century.

The Three Laws were born in a conversation between Asimov and his editor, John W. Campbell, in the summer of 1941. At the time, robot stories were dominated by the “Frankenstein trope”: robots as rebellious, destructive creations that turn on their human creators,