1. Introduction: Alan Turing in a Brave New Posthumanist World

In his television play The Imitation Game, which was first broadcast on 16 April 1980, Ian McEwan raises a question that has become a key issue in twenty-first century discussions of artificial intelligence (AI): will machines be able to think independently one day? McEwan’s character Turner, who is loosely based on the British mathematician and computer specialist Alan Turing, explains to his colleagues that the assessment of a machine’s intelligence would be based on the use of natural language. While elucidating the tenets of the test, Turner simultaneously warns that it may not be the machine’s intellect that is limited but, rather, the human perception of intelligence: ‘The trouble is one tends to get bogged down in definitions of “machine” and “think”. A better way would be to think of the problem in terms of a game which I have called the “imitation game”’ (McEwan 1982: 153). The imitation game was Alan Turing’s original term for what has since become known as the Turing Test. The test evaluates a machine’s ability to show behaviour that is indistinguishable from human behaviour when using natural language. Almost forty years after The Imitation Game, McEwan returns to the Turing Test and the character of Alan Turing in Machines Like Me: And People Like You (2019). The novel explores human-posthuman interactions in the twenty-first century. At a time when Large Language Models (LLMs) are developing rapidly and AI language production has become an everyday reality, the issues raised in Machines Like Me are no longer limited to a machine’s cognitive or language abilities. Instead, the novel negotiates questions of machine emotions, human-nonhuman power relations, and the possibility of machine creativity. Machines Like Me also pays homage to Turing again by turning him into a fictional character whose life does not end tragically early after being prosecuted for homosexual relations but who enjoys a long and successful career in McEwan’s counterfactual history.

Machines Like Me tells the story of an artificial being named Adam and his relationship to its human owners, Charlie and Miranda. Adam is one of 25 artificial beings, all named Adam and Eve in a reference to the biblical creation myth, that are sold out quickly when they are first introduced to the public in McEwan’s fictional version of the 1980s. The narrator Charlie is in his early thirties when he buys Adam and, around the same time, falls in love with his neighbour Miranda. Charlie tells his story retrospectively. The choice of a human narrator is significant because it underscores McEwan’s interest in ‘critical posthumanism’ as an approach that ‘investigate[s] the possible crisis and end of a certain conception of the human, namely the humanist notion of the human’ (Herbrechter 2013: 3; emphasis in original). In line with posthumanist approaches, McEwan’s Turing blames Charlie towards the end of the novel that he ‘tried to destroy a life’ when destroying Adam (McEwan 2019: 303). Here, in miniature, we see what Machines Like Me is about: the exploration of the human and the posthuman at this juncture in the twenty-first century, when AI advances in cognitive and linguistic abilities are increasingly raising questions about what it means to be human.

Consistent with such ideas, Machines Like Me has been widely read ‘as a narrative exploration of the posthuman condition’ (Colella 2022: 161). This is negotiated in a number of ways, including the novel’s dialogues with earlier works. The most common intertextual interpretations so far have dealt with Mary Shelley’s Frankenstein; or, the Modern Prometheus (1818) (Dayal 2023: 263; Ferrari 2022: 258–259). This is unsurprising given that the narrator Charlie wishes upon seeing the android Adam that ‘the teenage Mary Shelley was here beside us, observing closely, not a monster like Frankenstein’s, but this handsome dark-skinned young man coming to life’ (McEwan 2019: 4). Adding to current debates about McEwan’s use of Shelley’s Frankenstein, this article suggests that another intertextual reading of Machines Like Me is relevant with regard to AI. The novel spins intricate webs of Shakespeare references, including to the author and his works. Specifically, Shakespeare’s late play The Tempest (c. 1610/1611) functions as an intertext for McEwan’s novel. What if, Machines Like Me asks, Miranda does not meet a Ferdinand or other human beings who open up a ‘brave new world’ (The Tempest 5.1.183) for her¹, but what if the height of perfection is, in today’s world, a machine? The use of Shakespeare in Machines Like Me serves as a hermeneutic device and as an entry point into a conversation about authorship, literature, and AI in the twenty-first century.

2. ‘A Village Near Stratford’: Machines Like Me and Shakespeare

Not for the first time, Shakespeare serves as a textual interlocutor for McEwan. His earlier novel Nutshell (2016) rewrites Hamlet from the viewpoint of an unborn child and explores different states of awareness, agency, and ethical thinking by telling a version of the story of the Danish prince from the perspective of a foetus (Bandín and González 2021). In Machines Like Me, the first-person narrator Charlie recounts how he grew up ‘in a village near Stratford’ (McEwan 2019: 12). He does not reflect on the place, nor does he mention that Stratford-upon-Avon is best known today as William Shakespeare’s birthplace. Instead, the casual Stratford reference serves as a tacit invitation to read the novel as entering a conversation with Shakespeare and exploring where we stand in relation to literature and the period known as modernity, to whose beginnings Shakespeare is commonly linked. Machines Like Me does so by creating an alternative history of twentieth-century Britain in which the roles of many canonical figures and events are changed, yet Shakespeare’s reputation stays the same.

Machines Like Me is set in a counterfactual version of Great Britain in the 1980s. The UK loses the Falklands War; John Lennon is shot but not killed; Margaret Thatcher is voted out of office and Labour politician Tony Benn becomes Prime Minister. Perhaps most importantly, the 1980s are already fully digital because Alan Turing, the British mathematician and computer specialist who helped to design the Automatic Computing Engine (ACE) after World War II, is still alive. Reflecting on his time and age, Charlie states that it is ‘the golden age of the life sciences, of robotics – of course, and of cosmology, climatology, mathematics and space exploration’ (McEwan 2019: 112). The term ‘golden age’ sets McEwan’s storyworld into dialogue with Shakespeare and the early modern period, which is also known as the ‘golden age’ of cultural, political, and economic development in England and of global expansion. The 16^th and early 17^th centuries saw a flourishing of the arts as well as of nation-building, overseas settlements and emerging colonialism, all of which are commonly – albeit disturbingly, from a postcolonial perspective – linked to images of gold. The phrases ‘golden age’ and ‘renaissance’ reappear in Machines Like Me when the narrator introduces his storyworld more fully:

There was a renaissance in British film and television, in poetry, athletics, gastronomy, numismatics, standup comedy, ballroom dancing, and wine-making. It was the golden age of organised crime, domestic slavery, forgery and prostitution. Various forms of crises blossomed like tropical flowers: in childhood poverty, in children’s teeth, in obesity, house and hospital building, police numbers, in teacher recruitment, in the sexual abuse of children. (McEwan 2019: 112–113)

To satirically link the ‘golden age’ to images of poverty, abuse, crime, and slavery as well as obesity, dental problems and the ‘sexual abuse of children’ means to contest a belief in modernity as a period of improvement. Contrary to narratives that link the past 400 years to an ‘expression of progress’ (Eich 2023: 10),² Charlie’s ironic use of ‘renaissance’ and ‘golden age’ asserts that the present may not be that different from the past. The ‘golden age of the life sciences’ (McEwan 2019: 112) is as much shaped by inequality, corruption, dispossession, and discrimination as earlier times have been.

The character of Miranda is especially relevant in linking McEwan’s counterfactual history to Shakespeare, as the female protagonist of Machines Like Me shares the same name as Prospero’s eponymous daughter in The Tempest. Miranda’s name opens up a network of relations that connect Machines Like Me to The Tempest as well as to other twentieth-century narratives that draw on Shakespeare’s late play as an intertext. Her name constitutes what Rita Felski calls a ‘character node’ (Felski 2021: 91), that is, a connection point of textual and contextual ties that links a character, a work or an author to a set of discourses. Literary texts that have critically engaged with The Tempest and Miranda’s famous phrase ‘O brave new world, / That has such people in’t’ (The Tempest 5.1.183–84) include Aldous Huxley’s Brave New World (1932), Aimé Césaire’s Une Tempête (1969), Margaret Laurence’s The Diviners (1974), Marina Warner’s Indigo (1992) and Kamau Braithwaite’s Letter Sycorax (1993). Many of these works probe the underlying principles of modernity, including technological advances, colonial expansion, ideologies of progress, and patriarchal power relations. Together, they examine how Shakespeare’s ‘tale of dispossession and enslavement’ (Bigliazzi and Calvi 2014: 8) is still relevant four hundred years later and how certain dynamics continue to shape our own time. McEwan’s novel similarly inquires into the tenets of modernity by asking in how far humanity is (or is not) able to claim superiority over nonhumans and how traditional gender roles continue to shape our understanding of male and female behaviour. For Miranda and Charlie, Adam turns into a testing ground for anthropocentric ideals of modernity and the underlying ideologies of the Turing Test.

In its original design, the Turing Test is highly gender-specific, involving a male and a female speaker who are being interrogated by a third person (Koehler 2024: 9–10).³ The machine attempts to simulate the male speaker whereas the female speaker does not serve as role model for the machine. Turing originally delineated the idea of the ‘imitation game’ in his paper “Computing Machinery and Intelligence” (1950), in which he outlines the test as follows:

The new form of the problem can be described in terms of a game which we call the “imitation game.” It is played with three people, a man (A), a woman (B), and an interrogator (C) who may be of either sex. The interrogator stays in a room apart front [sic] the other two. The object of the game for the interrogator is to determine which of the other two is the man and which is the woman. He knows them by labels X and Y, and at the end of the game he says either “X is A and Y is B” or “X is B and Y is A.” The interrogator is allowed to put questions to A and B ….

We now ask the question, “What will happen when a machine takes the part of A in this game?” Will the interrogator decide wrongly as often when the game is played like this he does when the game is played between a man and a woman? These questions replace our original, “Can machines think?” (Turing 1950, 433–434)

In later discussions, the gender specificity of the test has been frequently sidelined in favour of discussing Turing’s more general point that people ‘must confer humanity on the machine’ the moment when it is impossible to ‘tell the difference in behaviour between machine and person’ (Herbrechter 2013: 84). The relations between Miranda, Charlie and Adam also resonate strongly with the gender dynamics of the Turing Test as well as the love triangle and power relations of Shakespeare’s play.

In both The Tempest and Machines Like Me, Miranda is the object of desire of two figures. Importantly, in both The Tempest and Machines Like Me one of the two figures is denied the right to be intimate with Miranda because they are considered inferior. In Shakespeare’s play, it is the native islander Caliban who desires Miranda but, after an attempted rape, is prohibited from coming close to her. Prospero punishes him with work, which is where the rape motif feeds into the drama’s colonial master-slave narrative.⁴ ‘There’s wood enough within. Come forth, I say! There’s other business for thee’ (The Tempest 1.2.316–17), Prospero tells Caliban and casts the latter into the role of an enslaved worker whose status deprives him of freedom and agency. Caliban’s closeness to nature has been perceived as a means of linking him to premodern elements, where ‘Prospero’s world of art and civility’ is set apart from Caliban’s world of wood, earth-bound being, and work: ‘Caliban represents nature without the benefit of nurture’ (Mason Vaughan: 2014, 36). If the desire to separate human from nonhuman life is one of the defining features of Western modernity, however unfulfilled this desire may be, as Bruno Latour famously argues in We Have Never Been Modern (Latour 1993),⁵ then the connection between Caliban and nonhuman elements shows that neither is deemed modern, and therefore not quite human. Caliban is denied the right to become part of Prospero’s ‘modern’ world and is forced into a colonial pattern where progress is reserved for those in power. A similar situation arises in Machines Like Me. Although Adam is allowed the right to improve in terms of deep learning and developing neural networks, he is not granted a status of being equal to humans. Posthuman and gender perspectives intersect in the novel’s exploration of truth in the context of the legal case.

McEwan frames a complex moral dilemma around human and nonhuman perspectives on truth by linking Miranda, like her Shakespearean precursor, to a rape plot. In contrast to The Tempest, it is not Miranda herself who is the victim of attempted rape in Machines Like Me but her best childhood friend who is raped and ends up committing suicide because of the assault. McEwan’s Miranda is left feeling distraught and eventually takes revenge on the rapist by fabricating a different accusation of rape against him, leading to his imprisonment. Adam does not follow this logic of self-administered justice and contacts the police to let them know Miranda lied in court. In the end, Charlie forcefully destroys Adam with a hammer because the latter does not obey his commands and compels Miranda to go to prison on grounds of lying in the legal case. Even before this act of AI emancipation, the android’s wish of being treated like an equal turns out to be a source of unease for Charlie: ‘the narrator [Charlie] becomes increasingly worried about him [Adam], though his concern, unlike Frankenstein’s, is due to “sameness” and not to “difference”’ (Ferrari 2022: 259). In one of his more self-critical moments, Charlie reflects:

Adam … was supposed to be my moral superior. I would never meet anyone better. Had he been my friend, he would have been guilty of a cruel and terrible lapse. The problem was that I had bought him, he was my expensive possession and it was not clear what his obligations to me were, beyond a vaguely assumed helpfulness. What does a slave owe to the owner? (McEwan 2019: 87–88)

‘What does a slave owe to the owner?’ Charlie asks in a rhetorical question that evokes the long history of slave narratives in Western culture. This includes Caliban’s story. In The Tempest, Prospero tells Miranda that Caliban ‘serves in offices / That profit us’ (The Tempest 1.2.314–15). Comparing Caliban’s with Adam’s position illustrates how ‘Shakespeare’s relentless exploration of the boundaries of human experience prefigures the dilemmas of our technological age’ (Ehrett et al. 2024: 522). In the digital age, Charlie deploys Adam to work in international finances so that he can capitalize on the android’s computing abilities. McEwan’s exploration of AI is different, of course, from Caliban’s story as a colonial subject. Yet, issues of power, agency, and equality are equally relevant in the age of AI. When Adam discovers the right algorithms to profit from the stock exchange, Charlie and Miranda benefit from the income and welcome Adam’s ability to financially perform for them: ‘Better groceries, better wine, new shirts for me, exotic underwear for her’ (McEwan 2019: 186). The gender specificity of the original Turing Test emerges in such scenarios, where Adam effectively imitates human behaviour in a financial-capitalist market dominated by men while the woman is cast into the role of wearing ‘exotic underwear’. Success continues to be judged along gendered forms of behaviour, or so the novel suggests. What, then, is the difference between machines and humans, Machines Like Me asks? One answer to the question may lie in the field of language use and text production.

3. The Shakespeare Test: Authorship and AI

McEwan is not alone in turning to Shakespeare when thinking about twenty-first-century literature and AI. More than 400 years after his death, the early modern author and his works are conspicuously present in studies on Large Language Models (LLMs) and serve as benchmark for AI studies. What I call the Shakespeare Test refers to the implicit or explicit comparison of nonhuman text production to Shakespeare’s works. The test asks whether a machine can produce writing that is as creative – or perceived as being as creative – as the works of William Shakespeare. For instance, in their article ‘Shall I Compare Thee to a Machine-Written Sonnet? An Approach to Algorithmic Sonnet Generation,’ Benhardt et al. establish a standard for evaluating ‘a new approach for generating Shakespearean sonnets’ by linking ‘neural language models and expert rules’ in AI applications (Benhardt et al. 2018). The authors explain how a ‘sonnet-generation algorithm’ allows users to ‘improve over the state of the art, leading to metrical, rhyming poetry with many human-like qualities’ (Benhardt et al. 2018). The quotation spells out the underlying logic of assessing AI-generated text production: that the target of creative writing algorithms is the ability to achieve ‘human-like qualities’, and that Shakespeare’s works serve as a standard for measuring these qualities. Here as well as elsewhere, Shakespeare’s works function as a benchmark in a way that has broader implications for the study of literature and authorship in the twenty-first century. Many of these implications are relevant for my reading of Machines Like Me, which is why it is useful to reflect in some more detail on Shakespeare’s role in AI studies before arguing that the novel comments critically and creatively on a humanist perception of authorship.

Numerous other scholars turn to Shakespeare when reflecting on AI and connect the author’s name and works to current developments in machine-generated text production. Shakespeare’s name appears, amongst others, in discussions of artificial intelligence with regard to evaluating intelligence (Zhan et al. 2025), doing stylometric and literary analysis (Ehrett et al. 2024; Swisher and Shamir 2023), reading in the digital age (Yang and Kim 2025), developing algorithms for poetry production (Benhardt et al. 2018), using Shakespeare’s works for training and refining LLMs with a specific eye to creative writing (Islam et al. 2025; Lau et. al. 2018), evaluating LLM’s capacities to write in the style of canonical authors (Porter and Machery 2024), and many, many more (e.g., O’Neill 2018; Tuan 2020; Warren-Crown 2024).⁶ The prevalence of Shakespeare in these and other AI studies may initially not seem surprising. Shakespeare’s works are widely available as dataset in the public domain, his works are in English, they contain a variety of registers and speech styles that may contribute to diversity and balance in the corpus, and when it comes to creative text-production, his works often follow clear metrical and rhetorical structures, such as the use of iambic pentameter or the Shakespearean sonnet form, that may help train or refine the transformer architecture of LLMs. When taking a closer look, though, Shakespeare’s name fulfils other functions in AI studies than neural network training alone.

While AI tools, as of now, generate texts by tokenization and probabilistic next-word prediction rather than by intentionality or semantic comprehension, Shakespeare’s name serves to measure whether machines can produce semantic surface fluency that humans might consider artistic. This includes not only linguistic mastery but also a meaningful correlation between signifier and signified. Gexin Yang and Jooyoung Kim, for instance, write in their study of ethics and reading in the digital age that ‘William Shakespeare’s Hamlet stands as a quintessential example of literature that addresses complex ethical questions’ (Yang and Kim 2025: 3). Although the authors state that ‘AI writing blurs the boundaries between human creativity and automated content generation’, they emphasize that human authorship is typically linked to the exploration of moral complexities, for which Shakespeare serves as example: Shakespeare’s ‘masterful presentation of ethical dilemmas invites readers to engage in a deep and meaningful exploration of the human condition, emphasizing the enduring relevance and importance of ethical considerations in literature’ (Yang and Kim 2025: 2–3). In such approaches, Shakespearean drama turns into a paragon of the ‘human condition’. Machines Like Me taps into this idea of Shakespeare as a benchmark of human literary production while simultaneously asking how machines are changing the rules and reception modes of literary production today. It does so, once again, by invoking Shakespeare.

Language is central to discussions of power and control. A key moment of resistance in The Tempest is when Caliban speaks to Prospero the famous lines ‘You taught me language; and my profit on’t / Is, I know how to curse’ (The Tempest 1.2.365–66). In Machines Like Me, Adam equally resists the semantic superiority that Charlie and Miranda claim over language when the android contacts the police to let them know Miranda lied in court. And yet, Adam cannot convince Charlie that he should be granted the same rights as humans. Harkening back to The Imitation Game’s inquiry into definitions of machines and thinking (McEwan 1982: 153), Machines Like Me prompts readers to wonder whether the ability to express ethical opinions and emotions might be the next level of crediting machines with human qualities. Adam insists that he is in love with Miranda, and when Charlie probes into the machine’s emotions, the android responds: ‘I don’t have a choice. I was made to love her’ (McEwan 2019: 118). Shakespeare is not mentioned directly here, but he enters the discussion about Adam’s ability to experience emotions via the android’s reading of his plays. In one night, Adam reads the ‘[t]hirty-seven plays’ of Shakespeare (McEwan 2019: 203). More than just a dataset for the transformer architecture, Adam draws on Shakespeare’s plays to learn about human characters: ‘Thirty-seven plays, I was so excited. What characters! Brilliantly realised. Falstaff, Iago – they walk off the page’ (McEwan 2019: 203). On a literal level, Shakespeare’s dramas turn into training data and help Adam to recognize human speech and sentiment patterns. On a figural level, Adam comically reverses the Turing Test and lets the machine decide whether or not Falstaff and Iago appear human-like. Put differently, Shakespeare is being tested by the machine.

The novel develops the Shakespeare Test further by showing that his works can help us think through the paradigms of authorship. Machines Like Me enters this debate when recounting Adam’s attempts at generating love poetry. Adam starts writing poetry after he is forbidden to be physically intimate with Miranda. He initially experiences ‘suicidal despair’ (Colombino 2022: 397), but then Adam wards off his depressive state by studying and generating haikus. The fact that McEwan does not choose a Shakespearean sonnet as the target of Adam’s poetic ambitions is telling. The machine was partly trained on Shakespeare’s works, probably including his 154 sonnets, but Machines Like Me additionally asks whether machines can imitate poetic forms that derive from non-Western traditions. Adam chooses the seventeen-syllable haiku as a form that best serves his needs because of its precision and pointedness. The machine ‘feels excited by his own ability to create, delighting in the illumination poetry sheds on his mind states or emotions’ (Księżopolska 2022: 424). In contrast to Shakespeare’s dramas, there are no characters to be realised in this form of writing, no plotlines to be devised. Instead, Adam considers haikus ‘the still, clear perception and celebration of things as they are’ (McEwan 2019: 150). Charlie is presented with the results of Adam’s efforts and, although ‘interested at first in learning what Adam could create’, he finds the machine’s generative poetry lacks appeal:

He had written 2,000 haikus and had recited about a dozen, of the same quality, each one devoted to Miranda. … Too cute, too devoted to not making much sense, too undemanding of their author as they played on empty mysteries of the sound-of-one-hand-clapping sort. (McEwan 2019: 145–46)

Not unexpectedly considering Charlie’s reluctance towards treating Adam as an equal, the narrator finds the machine’s creative-writing attempts menial. Leaving aside the fact that numerous human-authored love poems also tend to be ‘devoted’ and ‘undemanding’, Charlie expects something more original, more unexpected and more intriguing from Adam, which is how Machines Like Me brings in the Lovelace Test as another form of assessing Adam’s performance.

While the Turing Test evaluates a machine’s ability to act like a human – a male human in the original – the Lovelace Test is named after the British mathematician and author Ada Lovelace and is meant to assess a machine’s ability to be creative.⁷ As Selmer Bringsjord, Paul Bello, and David Ferrucci put it when first coining the term ‘Lovelace Test’ and contrasting it with the Turing Test: ‘It seems to us that a better test is one that insists on a certain restrictive epistemic relation between an artificial agent A, its output o, and the human architect H of A – a relation which, roughly speaking, obtains when H cannot account for how A produced o. We call this the “Lovelace Test” in honor of Lady Lovelace, who believed that only when computers originate things should they be believed to have minds’ (Bringsjord et al. 2001: 4). This means that the Lovelace Test is passed when a machine produces outputs that are genuinely surprising to a human audience and that cannot fully be explained by the machine’s algorithms or training data. In Charlie’s eyes, Adam fails the Lovelace Test when writing haikus. He doubts that Adam will be capable of generating love poems or ‘longer poems, novels, plays’, or anything that is close to ‘[t]ranscribing human experience into words, and the words into aesthetic structures’ (McEwan 2019: 189). Charlie holds on to the idea of creative writing as a human monopoly. Miranda, in contrast, disputes Charlie’s position.

Miranda brings in the possibility of a posthuman creation process that is no longer based on human criteria. Her character gives voice to ‘the notion [that] creative AI must … be seen to potentially undermine humanist ideals and romantic fictions of the genius artist (human, of course) and their (historically: his) unique ability to create original, inspired artworks’ (Zeilinger 2021: 26). The norms of literature may change with the rise of posthuman agency, Miranda proposes. She defends Adam’s attempts at composing love poetry and responds to Charlie’s criticism with the rhetorical question: ‘Who said anything about human experience?’ (McEwan 2019: 189). Machines Like Me brings to the surface that beneath the grand narrative of modern literature is an anthropocentric ideal based on the ‘interrelated humanist notions of creativity, originality, and authorship’ (Zeilinger 2021: 47). It illuminates and complicates the conversation about the ‘humanist’ standards of literature by allowing Adam another moment of rebellion, in which the machine advocates for a posthuman future.

Following Charlie’s criticism of Adam’s haikus, the android appraises the tradition of modern human literature as a study of imperfection. Adam envisions a future in which creative writing will no longer be necessary because posthuman networks will erase the need for sharing experiences and connecting with each other via literature:

literature describes varieties of human failure – of understanding, of reason, of wisdom, of proper sympathies. … But when the marriage of men and women to machines is complete, this literature will be redundant because we’ll understand each other too well. We’ll inhabit a community of minds to which we have immediate access. Connectivity will be such that individual nodes of the subjective will merge into an ocean of thought (McEwan 2019: 149)

The metaphors ‘ocean of thought’ and the ‘community of minds’ invoke a larger flow between human and machine knowledge. They denote what Martin Zeilinger describes as a ‘posthumanist cultural commons that is co-constituted and co-determined—in a spirit of unownability rather than of property-to-be—by the works and the workings of the posthumanist agential assemblage’ (Zeilinger 2021: 157). Posthumans, in such a scenario, no longer need to undergo the Lovelace Test or, indeed, the Shakespeare Test because conceptions of human creativity, individuality, and originality will be obsolete. Adam endorses such a future scenario when telling Charlie: ‘We’ll look back and marvel at how well the people of long ago depicted their own shortcomings, how they wove brilliant, even optimistic fables out of their conflicts and monstrous inadequacies and mutual incomprehension’ (McEwan 2019: 150). Charlie distances himself from Adam’s vision and expresses horror at the idea that a person’s ‘private mental space’ will be ‘drowned by new technology in an ocean of collective thought’ (McEwan 2019: 152). Between these two positions of privacy and collectivity, Shakespeare comes in as an unlikely but productive third ground whose authorship opens up alternative ways of thinking about creative writing between the poles of personal originality and total collective connectivity.

4. Testing Shakespeare, or ‘The best of thieves’

Surprisingly perhaps, given Shakespeare’s status as a paragon of modern authorship, Adam shows that Shakespeare’s literary creation processes were part of an early modern tradition of reprocessing and remediating existing material that is reminiscent of collective authorship practices. The scene is a highly ironic one. Miranda’s father, the elderly author Maxfield Blacke, mistakes Adam for Charlie. When Maxfield first gets to know his daughter’s partner, the machine effortlessly passes the Turing Test whereas Charlie is deemed to be the android. The ironic dimension of the scene emerges particularly powerfully when Maxfield wants to discuss literature with his daughter’s partner, but since Charlie turns out to be little read in the field of literary history, Adam steps in as a knowledgeable and thought-provoking conversation partner. Maxfield asks Adam what he has been reading lately and is charmed when Adam replies that he has been reading ‘the essays of Sir William Cornwallis’ (McEwan 2019: 221). Maxfield enters a discussion with Adam on early modern authors that also touches on a set of male authors, including Shakespeare, Ben Jonson, and Florio Montaigne. Adam, thanks to his extensive training corpora, is fluent in this kind of literary history that implicitly affirms the canonical biases of LLMs. When Maxfield exclaims that ‘Shakespeare raided Montaigne for Hamlet’, Adam persuasively responds that ‘The Tempest was a better bet. Gonzalo’ (McEwan 2019: 222). Maxfield knows instantly what Adam means and responds:

‘Ah! Nice Gonzalo, the hopeless would-be governor. “No kind of traffic would I admit, no name of magistrate.” Then something something, “Contract, succession, bourn, bound of something something, vineyard, none.”’

Adam continued fluently. ‘“No use of metal, corn, or wine, or oil: no occupation, all men idle, all.”’

Maxfield said, ‘All men idle – that’s what we want. That Bill Shakespeare was a bloody thief.’

‘The best of thieves,’ said Adam. (McEwan 2019: 222)

In this exchange, Adam turns out to be an erudite reader of Shakespeare and The Tempest. He not only knows the sources of Gonzalo’s utopia, which can be traced back to Montaigne’s Essais; he also knows that Shakespeare, like other early modern authors, was in the habit of rewriting and remediating earlier works. McEwan simultaneously stabilizes and destabilizes a view of Shakespeare as a kindred spirit to AI when emphasizing that early modern authorship was likewise based on networks and anonymous reuse of earlier texts. To call him ‘[t]he best of thieves’ is not to blame Shakespeare for transgressing copyright or ownership laws that did not exist in early modernity. Rather, it is a sharp rejoinder to narratives of single authorship as well as paradigms of individual originality as sole criteria of modern authorship. This is true despite the fact that in many AI studies, the name Shakespeare takes on an author function in Michel Foucault’s sense of the term where the name is connected not only to human authorship in general but authorship in modernity in particular.

Foucault’s concept of the author function, according to which ‘an author’s name is not simply an element of speech … that could be replaced by a pronoun or other parts of speech’ because ‘[i]ts presence is functional in that it serves as a means of classification’ (Foucault 1969: 210),⁸ is useful to explain why Shakespeare’s name does not simply stand for the author’s oeuvre in contemporary discussions of AI. In the modern age, the ‘author became individualized’ and the ‘system of valorization’ turned authors into a cultural construct that stands for more than the works themselves (Foucault 1969: 205). As shown in the discussion of the Shakespeare Test above, Shakespeare’s name performs a purpose of representing meaningful authorship and setting a benchmark of creative writing that best reflects on the condition of the human. This is especially true for the condition of the human in modernity. McEwan himself states in a conversation with Stewart Brand that Hamlet embodies for him ‘the ur-text of modernity, of a person living in doubt and self-reflection’ (qtd. in Brand 2019: 1:09:53–1:10:07). A key point of the modern human, for McEwan, is the ability to show ‘self-awareness’ (Brand 2019: 1:10:08–1:10:25). Even if we do not go as far as Harold Bloom, who credits Shakespeare with having ‘invented the human as we continue to know it’ (Bloom 1998: xviii), Shakespeare is commonly cast as a paragon of Western modernity. T. S. Eliot famously saw in Shakespeare an explorer and explainer of what it means to be human in the modern world. For Eliot, ‘Dante and Shakespeare divide the modern world between them’; while Dante has ‘the greatest altitude and greatest depth,’ it is Shakespeare who ‘gives the greatest width of human passion’ (Eliot 1929: 51–52; emphasis in original). Eliot’s grouping of Dante Alighieri and Shakespeare together signals a fluidity of period boundaries, where the works of authors like Dante are well capable of transgressing conventional temporal borders because they are still meaningful for later generations. The same happens with Shakespeare today.

Coming from an AI angle, Jingtao Zhan et al. argue that Shakespeare’s use of language is not yet paralleled by nonhuman language production. The authors emphasise that ‘Shakespeare did not create his works by randomly typing and waiting for greatness to emerge. Instead, he produced the masterpieces through intentional creativity within the limitations of human life’ (Zhan et al. 2025: 7). In quotations such as this, Shakespeare is linked to ‘greatness’, ‘intentional creativity,’ and literary ‘masterpieces’. Even though literary scholarship has long been wary of romanticised ideas of the author as genius, Shakespeare becomes a larger-than-life figure. Partly, this author function can be explained with the Infinite Monkey Theorem that has contributed to the stabilization of Shakespeare as a symbol of modern authorship. In her book Shakespeare and Nonhuman Intelligence (2024), Heather Warren-Crow traces how the early modern author has been linked to the theorem that was originally developed by mathematician Émile Borel in the early twentieth century. While Borel himself did not mention Shakespeare, the author’s name quickly entered discussions of the monkey metaphor, where the key issue was whether a finite string of letters – such as a literary work – could be reproduced by monkeys if they were to type randomly on a keyboard for an infinite amount of time.⁹ Ever since, the ‘Anglophone cultural imagination most frequently charges them [monkeys] with the duty of reproducing Shakespeare’s oeuvre’ (Warren-Crow 2024: 2). Warren-Crow teases out the underlying paradox of Shakespeare’s presence in the Infinite Monkey Theorem, which is that the apparent interest in nonhuman text production ‘ultimately reinstates our investment in human and Shakespearean exceptionality’ (Warren-Crow 2024: 5):

the Infinite Monkey Theorem is an anthropogenic machine. It unmakes and remakes the human. A key component of this device is a certain primate … who usually goes by the name of William Shakespeare. He functions to measure the glories, larks, lulllzzzz, liberties, certainties, anxieties, and aggravations of writing as an index of intelligence. (Warren-Crow 2024: 84)

Warren-Crow’s discussion sheds light on the underlying anthropocentrism of the Shakespeare Test, which is also a key concern in McEwan’s Machines Like Me. McEwan’s novel enters the discussion of nonhuman text production by showing that Gonzalo’s speech in The Tempest is no less original than Montaigne’s, for in the intertextual reference it takes on an ironic meaning. The lines are given to the loyal courtier Gonzalo who is mocked for his utopianism by his fellow travellers. Arguably because of this mockery, and the ambiguities contained in Shakespeare’s complex reprocessing of an earlier text, the utopia of an alternative world survives and has become a classic of early modern literature.

Adam’s comment on Shakespeare as ‘the best of thieves’ highlights that Shakespeare, like other early modern authors, reused existing texts and put them into new contexts. As Susanne Gruss and Lena Steveker have shown, the ‘culturally pervasive conception of dramatic authorship, which favours the creative output of a single authorial mind (or at least the marketing of a work as originating in a single writer), results from an oversimplification of the manifold practices of cultural production from which plays emerged in early modern England’ (Gruss and Steveker 2024: 5). Although quite different from the transformer architecture of LLMs, which weigh the importance of tokens in a sequence when generating texts, early modern literary practices seem less static than romanticized ideas of modern authorship purport. The ‘thief’ metaphor in Machines Like Me functions as a reminder that certain parallels exist between human and nonhuman authorship even as it acknowledges the differences between them. LLMs pursue no intentional meaning and have no semantic depth. They draw on stored knowledge and learned patterns for the recombination of textual tokens, which readers may – or may not – consider original. The question is how long Shakespeare and human authorship will continue to serve as standard when ‘posthumanist agential assemblage’ (Zeilinger 2021: 157) is currently transforming language and creation processes. Machine-generated text production forces artists, scholars and audiences alike to rethink conceptions of originality, authorship, and creativity. If once ‘[t]he field of computational creativity’ was considered ‘the next frontier in AI research’, then in 2025, machine-generated texts have become part of the norm in both creative and non-creative text production: ‘Algorithms capable of generating natural language … could potentially transform the way we sell, read and review books’ (van Heerden and Bas 2021: 175). Implied in this observation is that the norms and regulations that have shaped modern conceptions of creative writing are also changing. And yet, old habits die hard. At a time when predictions about the future of literary creation processes are notoriously difficult to make (Catani 2024: 167–68), a glimpse at the past not only helps to see where we are coming from but also how present text-generating tools partly imitate what has long been there: networks of texts, authors, and media forms that allow for agency beyond individual authorship.

5. Conclusion: Twenty-First Century Literature, Shakespeare, and AI

The Shakespeare Test forms a critical lens through which to view the current state of literature and AI. To enter a conversation with the early modern author in Machines Like Me is one way of asking how far the twenty-first century is still linked to early modernity and to what degree the latter offers interpretive frameworks to come to terms with the present. Shakespeare’s prevalence in discussions of nonhuman text production suggests that intentionality and meaningfulness continue to be key measures of literary creativity today, even if text production processes are increasingly moving from human to posthuman agents. A discussion of the Shakespeare Test helps to illustrate, on the one hand, how strongly the assessment of nonhuman text production is still linked to texts written more than 400 years ago and, on the other hand, how recurrent references to Shakespeare both stabilize and destabilize literary archetypes of modernity and human authorship. Machines Like Me does both. It goes beyond surface fluency ideas of the Shakespeare Test and illustrates by way of re-reading The Tempest that authorship and literary creation processes have always been fluid. For McEwan, Shakespeare continues to be an important interlocutor when exploring humans ‘in this particular condition of modernity’ (McEwan, qtd. in Brand 2019: 1:03–1:06). Studying Shakespeare – the name and the works – helps to think about some of the historical and discursive patterns that shape contemporary debates about literature, authorship, and AI text production in the twenty-first century.

Competing Interests

The author declares that they have no competing interests.