Towards the Artificially Intelligent Playwright: Experimentation with Chat GPT
Dominic Chamayou-Douglas*
Abstract
The role of the playwright is threatened by the technological advance of language generator software, including Chat GPT. New technology may be incorporated into the playwriting process but can undermine the essentially human work of the writer. Although today, Chat GPT is incapable of writing a human-standard play unaided, both the nominal and expressive authenticity of any play may be challenged in the future due to the impact of language generator software. Perhaps rejecting the software’s use in playmaking is therefore the simplest solution in order to maintain the authenticity of the play and the playwrighting process.
Keywords: AI, Authenticity, Playwriting, Technology, Text
Chat GPT and the Theatre
Considering the topic of this essay, it is useful to define what Chat GPT is and only appropriate to ask Chat GPT to define itself.
Chat GPT, short for “Chat Generative Pre-trained Transformer,” is an advanced language model developed by OpenAI. It is based on the GPT architecture, specifically GPT-3.5, which is designed to generate human-like text responses in natural language conversations. Chat GPT is trained on a vast amount of text data from the internet, allowing it to understand and generate coherent and contextually relevant responses to user prompts.
Define Chat GPT
Leaving aside the question of whether the software is artificially intelligent, Chat GPT evidently excels in generating large amounts of text on almost any given topic far more quickly and efficiently than any human. The interface is designed to simulate a chatbot, with the user encouraged to employ a discursive manner by asking follow-up questions or “prompts” in order to refine or regenerate answers and to attempt to attain any specific objective progressively through repeated interactions with the software. The responses can be generated in a range of registers and tones to match the specifics of the request, as long as any contextual information which may inform the output has been provided.
To this end, GPT software has already been used within a theatrical production at the Young Vic in London in 2021. Titled simply AI, the part-performance, part-research piece, directed by Jenifer Tang, involved using GPT-3 (a previous version of the software) to develop an original playscript live in front of the audience, which was then instantly performed by a troupe of actors (Akbar). Each performance was thus an original, ephemeral event made possible by the power of the technology and the willingness of the audience to invest in the experimental nature of the work. The Czech company THEaiTRE has also used AI software (the exact program is not specified) to develop two pieces, Ai: When a Robot Writes a Play (2021) and Permeation (Czech Republic) (2022), which were performed as plays acknowledging, even accentuating, the involvement of AI in the writing process (Rosa et al.). It appears evident then, that AI software and specifically Chat GPT can be used to enable the generation of text for performance.
However, can a text, even when that text is thematically and linguistically coherent as well as being presented in a recognisable format and style of a play, be considered a play text without the intervention of a human? Both of the above examples involved human-computer collaboration to adapt the text for performance, presumably because without such amelioration offered by a human intermediary, the text is lacking in certain theatrical qualities and fails to conform to the rather nebulous idea of what constitutes a play. Perhaps the text is insufficiently original or interesting; lacking in intrigue, conflict or drama; unbelievable without being fantastical; derivative to the point of plagiarism or plainly offensive. Perhaps the text simply did not respond to the subjective taste of the primary decision-maker in the development process. Perhaps the text failed to generate a distinctive voice to pass “the test” proposed by Gary Whybrow, former literary manager of the Royal Court in London:
Just imagine taking a single page of a writer’s work and throwing it on the floor in a mass of other pages, written by other writers. If you can identify that writer from one page then they have a distinctive voice.
qtd. in Sierz 50
As Chat GPT is designed to provide a seemingly endless number of variations of a text with the style and vocabulary adapting to the human input, the software is practically incapable of possessing a voice but instead generates multiple distinguishable voices. Chat GPT is therefore incapable of possessing a unique voice, in the sense meant by Whybrow, as if the software continually recreates a voice for each text, it is impossible to identify the Chat GPT voice but only a multitude of Chat GPT voices. The decoupling of the relationship between the play text and the voice of the writer is central to the issue of whether Chat GPT is capable of producing an authentic play, as without a definitive unique voice, questions over authorship or whether the text is a play at all may be raised.
Authenticity in Art
Authenticity in art is commonly associated with the visual arts and the performance of music or theatre, with the term “authentic” possessing multiple meanings relating to different aspects of an object’s perceived authenticity. A common distinction found across the literature is drawn between concrete, verifiable facts concerning an object and more abstract concepts perhaps relating to the object’s design, manufacture or receival.
The UNESCO entry on the authenticity of cultural objects focuses on the credibility of sources relating to an object and how the cultural context critically informs an object’s value (22). The guidelines list multiple practical properties of an object that may be of relevance, along with “intangible heritage” and the “spirit and feeling” the object may invoke in others. Thierry Lenain similarly creates a binary distinction between the authenticity of an artist, which is a value of intrinsic nature relating to a person’s engagement, and that of an object, which is a property assigned through factual judgement, thus siting the imperceptible authenticity of an object with the maker and their intentions during the creative process (74). Denis Dutton also divides types of authenticity into two broad categories named nominal authenticity and expressive authenticity, a useful distinction with which to frame the discussion (259).
Nominal authenticity relates to the accurate identification of the origins, provenance and authorship of an object. Therefore, any issues of plagiarism, forgery, misidentification and misrepresentation fall within the category, as each issue is related to the problem of correctly naming the properties of the object (Dutton 260). Although it has been argued that forgeries and copies do not belong to the same class as original works due to the two possessing different attributes (Sagoff), these attributes are often not apparent to the casual observer, and thus our ability to correctly identify an artwork is fundamental to the construction of a canon or corpus. This in turn forms the basis for our shared notion of authenticated value within different art fields.
Expressive authenticity is connected to how the work of art expresses the nature and intention of the individual creator of the work, but also how the values of the collective or society are expressed through the artwork. Part of an artwork’s authenticity, then, is in how it relates to other artworks, cultural objects and ideas, as works of art are “manifestations of both individual and collective values, in virtually every conceivable relative weighting and combination” (Dutton 270). An artwork, including a play text, therefore, does not exist in isolation but forms part of a body of objects, of practice, which are interrelated and collectively express a common understanding of what constitutes that particular kind of artwork, while “this fact accounts for a large part, though not all, of our interest in works of art’”(270). Conversely, the creator may also be expected to be authentic to their personal authorial vision, which often requires ignoring or rejecting these established principles and common understanding in order to advance the artform within their field and create art that is considered original (Davies et al. 156).
Equally, for a work to be considered authentic, we may presume a degree of sincerity in the intentions of the author, with the authenticity of certain works intimately linked to the supposed sincerity of the creator (Neill). Expressive authenticity is thus multifaceted, context-dependent and prone to evolve, as society’s conception of how an artist and their art should demonstrate authenticity shifts with changing cultural values.
Our interest in an object is therefore often derived from the specific actions and practices that led to its creation, as well as the historical and cultural context which informed the creator during the process. The object in itself, devoid of the details of how it came into being, may become devoid of meaning or at least a communal held notion of meaning as it relates to the object’s value, quality and originality. Aesthetically, it may be argued that an object should in any case be evaluated solely on its appearance. Termed, the “appearance theory,” the argument supposes that if a faked object brings pleasure or satisfaction to the observer or reader who cannot distinguish between the genuine and the imitation, there are only moral, not aesthetic, grounds for preferring the former (Stalnaker 462). This position refutes the value of authenticity in art by inciting us to only consider that which we can directly experience or perceive with our senses. Such an argument may be especially useful to consider when judging a new form of art or artist in order to ensure any evaluation is as objective as possible, rather than being unfairly biased by comparisons with a pre-existing body of work.
The authenticity of computer-generated art can be framed in relation to Walter Benjamin’s concept of “aura,” a malleable concept connected to ritual, the location of production and the original use of the object (3). Benjamin argues that mechanical production erases an object’s aura due to the dislocation of the process from traditional methods of production. Following this argument, all computer-generated art must thus be lacking in aura, and several theorists have argued that computers are incapable of creating authentic art due to the lack of human agency, emotion and experience in the creative process (Boden; Ekárt et al.), as well as the absence of intention on the part of the computer (Huang and Sturm 2).
Another common critique is of a computer’s inability to create art which in not simply imitative and repetitive due to the algorithmic nature of a computer program’s basic design, which is based on reproducibility (Huang and Sturm 6). Alternatively, Mark Coeckelbergh highlights that judging computer-generated art alongside human art is unhelpful. As there is a general presumption that computer intelligence is trying to imitate human intelligence, which is not necessarily the case, the logical corollary of such a presumption is that the art generated should imitate or resemble human art. Coeckelbergh thus argues that computer-generated art should be judged separately on its own terms as a related but distinct form of art, rather than only being set in comparison to human art (296). This viewpoint echoes the “appearance theory” argument above by similarly questioning the usefulness of the concept of authenticity in relation to computer art altogether, as the criteria of authenticity excludes all forms of computer art. Authenticity is arguably a term that is only of use in application to human art, as only humans are capable of expressing the various forms of authenticity expected in the artistic process.
Authenticity and Chat GPT
Denis Dutton proposes that whenever the term authenticity is applied, “a good question to ask is, ‘Authentic as opposed to what?”’ (259). In this case, the comparison is between a play text written entirely by a human, without recourse to text generative software and one written either entirely or in part by Chat GPT. I am therefore distinguishing between Chat GPT and similar AI-powered language generator software (Google Bard, Jasper, Microsoft Bing, and so on) and all other writing software (Word processors, Document formatters, and so on). Similarly, a play written by a human with the help of Chat GPT, even if only to generate ideas, will be considered as AI aided and thus indistinct from a play written entirely by Chat GPT.
In terms of nominal authenticity, first, there is the question of how Chat GPT might cause issues of plagiarism, misrepresentation and misidentification, and secondly, whether the software is currently capable of being used in this capacity. By generating a play text, Chat GPT may potentially be accused of plagiarism given the technology’s method of drawing from data across a wide range of internet sources, an accusation already being levelled at AI-powered image-generating software (Wong). Plagiarism may be more directly apparent in the case where text has been introduced to the software along with instructions to imitate the work. These AI-generated works could then in theory be misidentified as belonging to a certain playwright or misrepresented as such for professional or personal gain. It should be highlighted that in the examples of the plays using Chat GPT cited above, there can be no issue of misrepresentation or misidentification as the theatre makers and producers were explicit and open in communicating how the technology has been used to produce the text.
To assess Chat GPT’s current capacity to be used in cases concerned with nominal authenticity, I prompted the software to generate plays in the style of well-known playwrights to assess whether the software is capable of imitating specific voices, as well as the structure and themes commonly associated with those playwrights. I focused on asking for plays in the styles of anglophone playwrights who have published a significant number of plays and have a distinguishable style across their work, perhaps possessing “a voice” referred to by Gary Whybrow (above), to ensure that enough material existed from which Chat GPT could draw and to allow for the greatest possibility of detecting a realistic imitation.
Prompted to write a play in the style of Tom Stoppard, Chat GPT produced a mini play of four very short scenes entitled “A Playful Entanglement,” along with detailed character descriptions and a short text explaining essential elements of Tom Stoppard plays, which the software has attempted to incorporate (“Tom Stoppard”). The length and structure do not resemble Stoppard’s plays which are far longer and often involve non-linear time structures, such as in Arcadia and Rock and Roll. The character descriptions are superficially similar to certain Stoppard characters. The Chat GPT generated “Alexander: A brilliant physicist, prone to philosophical musings” might be likened to the philosopher George Moore in Jumpers, while “Olivia: Charles’ wife, an aspiring playwright with a flair for the dramatic” seems to be a composite character from The Real Thing, a play about a troupe of actors and a playwright. The dialogue incorporates common Stoppard themes, such as finding real-life parallels for complex philosophical or scientific questions as in The Real Thing or Arcadia, but it bears little resemblance to genuine Stoppard text. For example, scene 2 is only four lines long and is heavily expositional.
Scene 2: Charles and Olivia’s Living Room
“Tom Stoppard”
(Charles and Olivia sit on a cozy couch, sipping tea. Sebastian bursts in, bursting with excitement.)
Sebastian: Charles! I’ve done it! I’ve solved the equation! The answer lies within the fabric of the universe!
Charles: (Teasingly) Ah, Sebastian, always the enthusiast. Do share your findings, my friend.
Sebastian: (Eagerly) It’s all about the interconnectedness of particles. They dance with each other, each step influencing the next. It’s like a grand cosmic tango!
Olivia: (Pensively) How fascinating! It reminds me of the intricacies of human relationships, where every action sets off a chain reaction.
The vocabulary and phrasing are recognisably an English of the upper middle classes, the class that Stoppard has often drawn his characters from, especially in a phrase such as “Do share your findings,” a formulation that indicates Charles is educated. Although the subject of conversation is solving a physics equation, the tone used is light and faintly humorous, almost flippant, an attempt to emulate Stoppard’s wordplay and comic asides. However, behind this simulacrum there is no idea, no plot, no concept. The words are strung together in comprehensible sentences, but the equation that Sebastien is referring to in the extract has no meaning within the world of the play, it is referred to once in this scene and then ignored thereafter. There is no world in which these characters exist and in which their conversations and actions can then construct and convey meaning. If we compare this extract to actual Stoppard text taken from The Hard Problem.
Spike Also, it’s not good science to call mother love a virtue, or even mother love.
13
Hilary You don’t think mother love is a virtue?
Spike You don’t call it a virtue, because at root its virtue consists in its utility.
Hilary Utility. Mother love?
Spike Genetically selected behaviour to maximise –
Hilary Spike, do you know anyone who believes that, really and truly?
Spike I don’t know anyone who doesn’t believe it. Parental behaviour. Hard-wired when we were roaming the savannah in small groups of hunter-gatherers. Mother and baby are in a cost-benefit competition. Have you ever seen a newborn infant screaming to be fed?—the anger—the noise—the face . . . ! The kid is laying it on, and it probably started in the womb.
This excerpt demonstrates many of the Stoppard traits common to much of his playwrighting: a philosophical or moral issue; linking of several concepts in rapid-fire dialogue; an undercurrent of conflict or disagreement; the contrast of a complicated argument communicated in colloquial language. Although superficial parallels can be drawn between the genuine and the imitation, there is an obvious difference in the tone of the discussion, the direction of the conversation and the desire of the writer to communicate understanding of the topic. Although Chat GPT is capable of pastiche, the software is visibly incapable of producing a credible imitation that could be misidentified or misrepresented as a play by Stoppard himself.
The other experiments trying to generate realistic imitations of other playwrights produced similar results, despite regular prompts to develop the style, the plot or the themes of the text in order to render a more accurate representation of the author’s work. The Arthur Miller imitations produced texts focusing on negative aspects of American capitalism, using plots concerned with working-class families, seemingly set in the mid-twentieth century (“Arthur Miller”). The Martin Crimp imitations, called Shadows of Solitude, used generally elliptical statements, had no discernible plot and avoided naturalistic setting (“Martin Crimp”).
Fading Shadows, the Caryl Churchill imitation, involves four female characters of different ages reflecting on the passing of childhood. The format recalls plays such as Far Away,while the dialogue is generally mundane with occasional emotionally-charged metaphors appearing (“Carol Churchill”) (links to all the outputs can be found in the Bibliography). Overall, the plays generated demonstrate the ability of Chat GPT to draw on the common themes, characters and writing styles of a playwright and use them as tropes verging on clichés. The software is evidently capable of parody which may be accurate enough to allow the original author to be identified through the imitation; however, the results generated, even with significant further prompting, are highly unlikely to ever pass as the work of one of the genuine playwrights. Not only are the plays thematically, linguistically and narratively underdeveloped, the results are more akin to pastiche than realistic forgeries.
In the case of misrepresentation, for a writer to claim to have written an entire play text generated by Chat GPT without acknowledging the fact seems categorically inauthentic. Even in the mildly absurd, but entirely possible case, of a writer using language-generating software to write a play text in their own style and then claiming the text to be their own creation, this certainly appears to undermine the nominal authenticity of the work. At least since Barthes’s canonical text, the idea that no author can claim to be at the origin of a text has been widely argued and acknowledged; however, using software which explicitly draws on other sources to generate text without passing through a human interlocutor, except to issue prompts, seems to belong to a different category than human written texts.
As for the question of expressive authenticity, I will endeavour to assess the texts with an “appearance theory” approach to establish whether the software is currently capable of producing a text that is recognisable as a play regardless of the method of production. Furthermore, I will not assume that this computer-generated text cannot be art as argued above, but I will instead examine whether the process of creating the text is so dissimilar to a human writing a text that the result should be considered as a separate category to human-written plays.
I began by asking the software to generate a social realist family drama, while employing a discursive approach as suggested by the developers (“Realist Play about a Family”). The play has recognisable characters, a dramatic situation concerning poor finances and identifiable classical dramaturgical tropes of characters suddenly entering with good or bad news to move the action forward. Much of the dialogue is expositional with the tone of the play moving excessively quickly as the characters’ emotions surge and dissipate from line to line. However, despite repeated prompts, the software seemed unable to develop the plot beyond the initial simple storyline of an uncle appearing to support his family financially. Instead, variations of the same scene are repeated, complete with the same pattern of characters entering and exiting and very similar dialogue being exchanged, first one week after the first scene, then one month, then three months and so on, with no change in the initial setup. Without an explicit command from the user to alter the fundamentals of the play being generated, the software appears to be caught in a loop, changing the dialogue but maintaining the initial narrative structure, characters and inter-relational dynamic. This form of repetition reads distinctly as non-human generated, as implicit in the request to write “the next scene” is the concept that the original scene must be developed in some way through the progression of the story or a change in form or style; however, Chat GPT seems to interpret the prompt as being further along in time rather than to generate a scene that might narratively follow the initial scene.
This characteristic of the text generated by Chat GPT could be identified in almost all of the different “chats,” regardless of the context or command. The generated plays tended to reiterate the initial premise rather than substantively evolve the material as a human writer may be expected to. Although repetition with minor variation is used by many playwrights as a dramaturgical tool, notably Martin Crimp and Caryl Churchill, the technique is employed intentionally to produce an effect or highlight an element of the text or action which is subtly changing. Chat GPT, however, repeats with variation despite prompts to do otherwise. Despite these reservations, the text generated is identifiably a play with common play aspects such as characters, a setting, a plot and a story arc, which could in theory be performed on a stage. However, the theatrical value of such an endeavour would be highly questionable.
The second experiment concerned the generation of a Greek tragedy using the dramatic structure of Aristotle’s Poetics (“Classical Greek Tragedy”).The texts generated contained character descriptions and scene-by-scene descriptions of the plots with no dialogue. All of the resulting texts were either direct replicas of pre-existing tragic plays, such as Oedipus Rex, or drew heavily on various recognisable Greek myths and characters. Even when the given prompt was for the generation of an original Greek drama, the text used existing characters from Greek mythology and a plot remarkably similar to the myth of Orpheus and Eurydice. Further prompts could eventually produce a more original outcome, but the results highlight the propensity of the software to plagiarise through the assimilation of multiple source texts. Perhaps the systematic mixing of elements from several myths with the adjustment of certain character names or places is not fundamentally different from the adaptation of a Greek myth by a human theatre maker who adopts a similar approach. However, the effect of observing the regeneration of text in a matter of seconds on a screen in an almost instantaneous reiterative process does not at least appear to be practically related to human activity. Although the end results produced by Chat GPT and a human, as in the words appearing on the page in a cohesive order with cogent references to readily attributable sources, may be cosmetically similar, the methods of production bear little resemblance to each other.
Most obviously, Chat GPT does not work on a human scale of time but generates texts so rapidly there is no time for internal reflection, external critique and incremental improvement, as occurs with most human-originated texts. This compression in the timespan of the production process is disconcerting to observe and engenders a sense of alienation towards the final output, which the user knows to be infinitely and instantly fungible. The fact that elements of the text may be vaguely recognisable while remaining clearly distinct from the sources being appropriated leads to dissonance in the user, caused by the difference between the text observable on screen and the prompt given to the software—that is, to create an original play.
Although the software does not claim to be producing an original text nor reproducing a recognisably human method of production, the creation of a play text is a process that, until now, has been reserved to humans. The fact that plays are closely related to our understanding of narrative and the importance of narrative to humans in terms of communication only enhances the sense that playwrighting is a priori a human activity. Therefore, the artificial generation of a specifically human activity, and the fact that the production process is observable in that the user can watch the words appearing on the screen, can produce an unnerving effect on the user.
Our communal understanding of playwriting as a process thus far makes no account for the explicit introduction of Chat GPT or equivalent at any stage of the production. The value of a play text is partly established through the perceived difficulty of writing one, the idea that another individual has laboured, physically, mentally and emotionally, to create a piece of writing that is capable of affecting readers and spectators. When we are experiencing a play, whether in public or private, the knowledge that the transmission of its ideas, in both form and content, has occurred from human to human is fundamental in our appreciation of the artwork. The ability to identify a specific individual to whom we can relate, at least in terms of our common humanity, is essential in society’s ability to judge and valorise the work itself. The reason that even minor works of Shakespeare are regularly revived at the expense of broadly equivalent works from the same period is the figure of Shakespeare himself. The idea that another human living hundreds of years ago wrote the words now being spoken before us, probably with a quill on a piece of paper or parchment, adds considerably to the cultural value of the play. A large part of the rationale for assembling in a room to watch another production of A Midsummer Night’s Dream is that the spectators present, and the wider society, place a high value on the importance of Shakespeare’s work and the ritual performance of work first seen several centuries earlier. The use of Chat GPT disrupts this fundamental sense of shared understanding as to how we can appreciate and evaluate play texts and their importance to our culture.
Even if the use of the software is limited to the generation of ideas, I would argue, from my experience with Chat GPT, that this usage is so disruptive to the process of playwriting that the question of whether the final text can be considered authentically a play must be posed. Each stage of the process of playwriting is valuable in the construction of a play in terms of its resemblance to our shared idea of a play, as well as imbuing the final product with embedded meaning. The initial starting point is critical in setting the path for the development process, providing both inspiration and a guide to the entire writing process. Often, the first question in interviews with artists about their work relates to the inspiration and the details behind the formative stages of the creative process. Therefore, to essentially delegate this stage of the work to a piece of software, fundamentally undermines a key element of the process and, therefore, the expressive authenticity of the final work itself.
Conclusion
The fact that Chat GPT is today incapable of producing a realistic forgery is irrelevant, as one day, perhaps sooner than we expect, the software may be able to do so with a speed and accuracy which we find shocking. In such circumstances, the nominal authenticity of any text will have to be questioned, as any writer could be effectively imitated by anyone else. One can imagine forgotten Elizabethan texts or unpublished works from the nineteenth century being discovered with surprising regularity. However, perhaps more disruptive will be to the collective notion of what it means to write a play if a writer, perhaps an established one with a reputation and distinct style, decides to make use of an AI-enhanced language generator and fails to acknowledge this support publicly. Photographers have already used AI to create prize-winning photographs, only acknowledging the use of AI in the process after the fact (Parshall). The day when an AI-aided text appears on a professional stage without an acknowledgment of the technology’s contribution seems to be inevitable.
If as a society we continue to value the play as an art form capable of entertaining us, critiquing society and enabling us to reflect publicly about ideas critical to the understanding of human nature, then we must similarly value the process which leads to the play’s creation. In the artworld, the replicable nature of many artworks has not always led to a devaluation in their worth; instead, it enhanced scrutiny of the authenticity of the work in question, as the furor over Leonardo Da Vinci’s Salvator Mundi demonstrates. Perhaps, as Mark Bly suggests, the two-step playmaking process of the playwright handing the text to a director and cast is outmoded, and the playwright must be integrated more fully into the active creation of the theatrical event (qtd. in Trencsényi), as this would evolve the role of the playwright outside of the capabilities of screen-based software and thus away from an easily programmable function that can be replicated by computers. However, I would argue that there is still a place for highlighting and venerating the authentic, purely human endeavour, of writing a play text, in a society which is never-endingly threatened by the looming progress of the artificial.
Bibliography
Akbar, Arifa. “Rise of the Robo-Drama: Young Vic Creates New Play Using Artificial Intelligence.” The Guardian, 2021. Accessed 26 June 2023.
Barthes, Roland. “The Death of the Author.” Aspen, vol. 5 and 6, 1967.
Benjamin, Walter. The Work of Art in the Age of Its Technological Reproducibility. 1936.
Boden, Margaret A. “Authenticity and Computer Ar.” Digital Creativity, vol. 18, no. 1, Mar. 2007, pp. 3–10, doi:10.1080/14626260701252285.
Chat GPT, and Dominic Chamayou-Douglas. “Define Chat GPT.” Open AI, 2023.
—. ‘Make up the Title of a Martin Crimp Play’. Open AI, 2023.
—. “Playwright’s ChatGPT Considerations.” Open AI, 2023.
—. “Write a Classical Greek Tragedy.” Open AI, 2023.
—. “Write a Play in the Style of Tom Stoppard.” Open AI, 2023.
—. “Write a Scene from a Realist Play about a Family .” Open AI, 2023.
—. “Write the Dialogue for Play about Loss of Childhood in a Spare Style like That Used by Carol Churchill.” Open AI, 2023.
—. “Write the Plot to a Stereotypical Arthur Miller Play.” Open AI, 2023.
Coeckelbergh, Mark. “Can Machines Create Art?” Philosophy and Technology, vol. 30, no. 3, 2017, pp. 285–303, doi:10.1007/s13347-016-0231-5.
Davies, Stephen, et al. A Companion to Aesthetics. John Wiley and Sons, Incorporated, 2009.
Dutton, Denis. “Authenticity in Art.” The Oxford Handbook of Aesthetics, edited by Jerrold Levinson, vol. 1, Oxford UP, 2009, doi:10.1093/oxfordhb/9780199279456.003.0014.
Ekárt, Anikó, et al., editors. Computational Intelligence in Music, Sound, Art and Design. Springer International Publishing, 2019, doi:10.1007/978-3-030-16667-0.
Neill, Alex in Gaskell, Ivan, and Salim Kemal. Performance and Authenticity in the Arts, 1999, doi:10.1017/CBO9780511520051.
Gaut, Berys, and Dominic Lopes, editors. The Routledge Companion to Aesthetics. Routledge, 2013, doi:10.4324/9780203813034.
Hern, Alex. “Elon Musk Joins Call for Pause in Creation of Giant AI “Digital Minds.’” The Guardian, 2023. Accessed 26 June 2023.
Huang, Ruijing, and Bob L.T. Sturm. “Reframing ‘Aura’: Authenticity in the Application of AI to Irish Traditional Music.” AI Music Creativity Conference, 2021.
Lenain, Thierry. “Authenticité Iconique vs Authenticité Artistique. à Partir d’une Œuvre de Francis Alÿs.” Noesis, no. 22–23, 2014, pp. 73–90, doi:10.4000/noesis.1888.
Parshall, Allison. “How This AI Image Won a Major Photography Competition.” Scientific American, 2023. Accessed 10 Oct. 2023.
Rosa, Rudolf, et al. “THEaiTRE: Artificial Intelligence to Write a Theatre Play.” AI4Narratives, 2020, pp. 9–13.
Sagoff, Mark. “The Aesthetic Status of Forgeries.” The Journal of Aesthetics and Art Criticism, vol. 35, no. 2, 1976.
Sierz, Aleks. Rewriting The Nation. 1st ed., Bloomsbury Methuen Drama, 2011, doi:10.5040/9781408166819.
Stoppard, Tom. “The Hard Problem.” The Hard Problem, Faber and Faber, 2019, doi:10.5040/9780571361991.00000004.
Trencsényi, Katalin. “‘Questioning Spirit’—Dramaturgy in America.” Critical Stages/Scènes Critiques, vol. 12, 2015.
UNESCO. Operational Guidelines for the Implementation of the World Heritage Convention. Intergovernmental Committee for The Protection Of The World Cultural And Natural Heritage, 2011.
Wong, Matteo. “Artists Are Losing the War Against AI.” The Atlantic, 2023. Accessed 12 Oct. 2023.
*Dominic Chamayou-Douglas is a PhD Researcher at the Universities of Kent (U.K.) and Lille (France) whose areas of interest include contemporary playwriting, narratology, translation theory and writing technology. He is a practicing playwright, theatre maker and translator whose work has been performed at various festivals and venues in the U.K. and France. He also co-founded and produced the Paris Fringe Festival, a multilingual international festival of contemporary theatre and performance.
Copyright © 2023 Dominic Chamayou-Douglas
Critical Stages/Scènes critiques e-ISSN:2409-7411
This work is licensed under the
Creative Commons Attribution International License CC BY-NC-ND 4.0.
