Artificial Intelligence and creativity in poetry: effect of AI-written poems on human emotions

6 September 2023 – Vol 1, Issue 1.

Abstract

In today’s modern age, we find ourselves confronted by an intriguing and debatable phenomenon: the profound achievements of the human intellect, manifest in the creation and propagation of Artificial Intelligence (AI), have also brought about some contentious outcomes. As we broaden AI’s application beyond purely technical tasks, we permit its infiltration into the realm of creative arts, traditionally a dominion of human emotion and perception. The field of poetry, no exception to this trend, is currently witnessing an accelerating exploration by AI, testing the waters of poetic creation. As we step into this brave new world of AI poets and algorithmic creativity, it inevitably makes us question whether we are prepared for such a shift and whether AI-generated poetry can invoke the same emotional resonance as that created by human hands.

Introduction

As we move further into the age of digital revolution, the expanding role of artificial intelligence (AI) in multiple dimensions of human life has become undeniable (Zhang, Zhu and Su, 2023). This expansion is not limited anymore to fields we traditionally consider technical or data-driven but has found its way into the creative realms as well. Today we can witness an increasing penetration of AI into painting, writing, photography, and, of course, an age-old artistic expression, poetry (Amabile, 2020).

Can AI generate poetry parallel to human quality? ‘William Blake and his Laptop’, by Gil Dekel and AI, from a portrait by Thomas Phillips (1807), 2023.

Historically, poetry has served as an intimate reflection of human senses. Through rhythm, word choice, and structure of their compositions poets channel their emotions, observations, and experiences. These elements, deeply rooted in human consciousness, facilitate a profound connection between the poet and the reader, touching the latter on physical, emotional, intellectual, and spiritual levels (Lund and Wang, 2023). Given this consideration, there emerges a pressing question: Can an AI, devoid of these intrinsic human experiences and emotions, craft poetry that resonates similarly with readers?

At the same time, looking at the capabilities of AI from another lens offers a contrasting perspective. AI’s inherent strength lies in its ability to rapidly analyze and draw from vast repositories of data (Aggarwal et al., 2023). In the context of poetry, this means that AI can scan through thousands of poetic works, spanning styles, eras, and emotions, to produce new compositions. Some argue that this extensive analysis might enable AI to generate poems that, while not born from human experience, might still resonate deeply with humans’ emotions (Demmer, Kühnapfel, Fingerhut and Pelowski, 2023). It might even offer insights and emotional depth that rivals traditional human compositions. The intersection of AI and poetry, therefore, presents a rich area of exploration, thus rising a dilemma, whether AI, while being deprived of an emotional component per se, can still stir the human soul in an equally evocative manner (Strehovec, 2023).

Considering the relative novelty of the topic of AI in general, there aren’t many comprehensive research on AI and poetry. Nevertheless, some research, such as Kirmani (2022), confirms that AI demonstrates an “astonishing level of control over language, syntax, stanza formation, rhyme, and meter” in its poetry (Kirmani, 2022, p. 576). However, when using it as a poetic or any other informational tool, users should consider that even the most contemporary AI platforms misinterpret facts from time to time. As Kirmani (2022) points out “bias often creeps in and is hard to control” (576). Further research supports the notion that AI-produced poems appeal to readers and can pass as human-written. Yet, the authors suggest that true creativity is rooted in the ability to innovate based on previous emotional experiences. Given that AI lacks an emotional component, it raises the question of whether AI-generated poetry can truly be considered creative (Köbis and Mossink, 2021).

The subject of this work, along with existing research, demands a deeper understanding of the subject of creativity as well. As creativity can be viewed from social, psychological, artistic, and even biological perspectives, it’s vital to comprehend its significance in relation to poetry.  Considering AI’s ability to elicit emotional responses in humans, as it will be further discussed, it’s essential to evaluate this capability in the context of established definitions of creativity.

Given the focus of this study, the existing literature review, and the persistent research gap in this field, in the current work we will try to answer the following questions:

1. 1. What impact do AI-written poems have on human emotions?
   2. Can AI-written poems be considered ‘creative’ based on prevailing definitions of creativity?

Literature review

Defining creativity

Creativity, in its quintessence, stands as a multifaceted, evocative phenomenon, composing a complex interplay of cognition, emotion, and innovation; however, pinpointing a single, universally accepted definition remains a challenging task. Historically, creativity was often reserved for discussions on divine inspiration or rare genius; the Romantic era, for instance, idolized the ‘tortured artist’, suggesting that true creativity stemmed from profound emotional turmoil. Yet, in contemporary discourse, creativity is perceived as a blend of different processes and abilities (Boden, 2004).

From a sociological standpoint, creativity isn’t merely an individual endeavour; it is intricately intertwined with society’s perceptions, values, and cultural milieu. Csikszentmihalyi (2015), in his seminal works, advocates for a systems perspective; he contends that creativity is not just the domain of the individual but is a triadic interrelation between the domain, field, and individual. A piece of art, be it a poem, painting, or musical composition, is judged creative not solely by its intrinsic merits but by its reception and endorsement within a particular socio-cultural context (Reckwitz, 2018).

Moreover, while these sociocultural definitions provide a macro lens, individual differences in creativity have also been a topic of great intrigue. Divergent thinking, often considered as a hallmark of creativity, is the ability to generate a plethora of potential solutions for open-ended problems. It is this very ability to break free from cognitive rigidity, to turn away from familiar paths, that differentiates creative individuals from their less creative counterparts (An, Song and Carr, 2016).

Biologically, the complexities of the human brain contribute significantly to our understanding of creativity. Recent neuroimaging studies have unveiled a complex neural group underpinning creative cognition; the default mode network, responsible for daydreaming and self-referential thought, juxtaposed against the executive control network, tasked with goal-oriented, focused thinking, collaboratively create a fertile ground for creativity. The juxtaposition and synchrony of these networks, neither working in isolation nor in strict tandem, echo the inherent paradox of creativity: structured yet free, focused yet divergent (Chatterjee and Vartanian, 2016).

Lastly, from a philosophical perspective, creativity has often been caught in debates concerning novelty and value. Is mere novelty, devoid of any tangible value, an emblem of creativity? Or does true creativity necessitate a symbiosis of the new and the valuable? These ruminations, although abstract, have profound implications for domains such as poetry, where the confluence of emotion, innovation, and linguistic prowess often blurs the border between novel and valuable, between the fleeting and the eternal (Klausen, 2010).

Connecting creativity, poetry and emotions

Historically, poetry has been considered as one of the most potent expressions of human creativity. This timeless form of literature uniquely captures the intricacies of human emotions, beliefs, and aspirations in ways that transcend the mere spoken word. Delving deeper into the nature of poetry, one discerns that it’s not just the words themselves that evoke emotions, but the alchemical combination of rhythm, imagery, and phonetic resonance. As a medium, poetry employs intricate language structures and metaphoric representations that strike chords with our intrinsic emotional landscapes, often leading to moments of epiphany or profound introspection (Obermeier et al., 2013).

In the broader discourse on creativity, poetry stands apart due to its inherent challenge of conveying complex emotions within often constrained formats. Whether it’s the brevity of a haiku or the rigorous structure of a sonnet, these limitations are paradoxically what drive the poet’s creativity. Restriction breeds innovation, pushing poets to delve deeper into their emotional reservoirs, crafting expressions that resonate with precision and depth. This amplifies the emotional impact on the reader, as each word, being meticulously chosen, carries a weight and profound intention (Johnson-Laird and Oatley, 2022).

Furthermore, the creative process in poetry is inherently intertwined with the poet’s own emotional journey. A poet doesn’t merely represent emotions; he undergoes them, as Johnson-Laird and Oatley (2022) point out.  This deep-seated connection between the creator and creation gives poetry its authenticity. When readers engage with such pieces, they’re not just connecting with the words but also with the myriad emotions that stand behind those words. This dual-layered connection is why poetry often feels more visceral and evocative compared to other literary forms.

Emotions, at their core, are deeply personal, yet universally relatable phenomena. Poetry, as a creative outlet, bridges the gap between these individual emotional experiences and the collective human psyche. Through creative expression in verses, personal feelings transform into shared experiences, leading to an emotional communion between the poet and the reader. This ability of poetry to encapsulate emotions and present them in relatable forms makes it a powerful tool for empathy, understanding, and human connection (Pițur and Miu, 2022).

These observations are also shared by Hilse, Griffiths and Corr (2007). Authors state that the therapeutic aspects of both reading and writing poetry cannot be overlooked. Engaging with poetry can be cathartic, allowing individuals to process, confront, and sometimes even transcend their emotional challenges. The act of writing, in particular, can be a path for emotional release, granting clarity and perspective. It’s no wonder that throughout history, during times of societal upheaval or personal conflict, poetry has often been the refuge for many seeking solace, understanding, or even a means of protest (Hilse, Griffiths and Corr, 2007).

Conclusively, the relationship between creativity, poetry, and emotion is symbiotic in nature. While creativity births poetry, it’s the deeply rooted emotional undercurrents that give it life and meaning. As readers, we’re drawn to poetry not just for its aesthetic or intellectual value, but for its uncanny ability to mirror our own emotional landscapes, validating, challenging, and enriching our human experience in the process (Pițur and Miu, 2022).

Impact of AI-written poetry on human emotions

Artificial Intelligence’s rapid arrival into the literary domain, especially poetry, has sparked a compelling discourse on the emotional resonance of AI-written compositions in contrast to their human-authored counterparts. Paul Ekman’s seminal work on basic human emotions delineates six core feelings: happiness, sadness, fear, disgust, anger, and surprise. These emotions, deeply ingrained in the human psyche, serve as a yardstick for evaluating the emotive depth of poetic pieces, irrespective of their origin (Ekman and Davidson, 1994).

Recent empirical studies shed light on how AI-generated poetry influences these core emotions. Preliminary findings suggest that, when readers are unaware of the authorship, AI-written poems can evoke emotions akin to those stirred by human-written pieces. For instance, AI-generated verses themed around melancholy or joy have shown potential in eliciting corresponding feelings of sadness or happiness in the readership, not remarkably different from traditional poems (Köbis and Mossink, 2021).

However, a nuanced divergence emerges upon deeper introspection. Human-authored poems, anchored in authentic experiences, often possess a raw emotional granularity — the subtle interplay of primary emotions with complex feelings like nostalgia, remorse, or anticipation. In comparison, AI compositions, though adept at mirroring the overt tones of happiness or fear, occasionally miss these intricate emotional undertones (Oksanen et al., 2023).

Moreover, the knowledge of a poem’s AI origin sometimes introduces a cognitive dissonance. Readers, when cognizant of the computational genesis of a poem, have reported a muted or altered emotional response, especially in emotions like surprise or disgust. This suggests that the perceived authenticity of experience, or lack thereof, can modulate emotional reception. (Wu, Mou, Li and Xu, 2020).

In conclusion, while AI-written poetry demonstrates a commendable capacity to engage with human emotions, its impact varies based on both its inherent linguistic structure and the reader’s awareness of its origin. The interplay between human emotions and AI-authored art remains a fascinating, multifaceted realm warranting deeper exploration.

AI-written poetry – is it truly creative?

The burgeoning field of AI-written poetry prompts a salient reevaluation of our foundational understanding of creativity. Traditional perspectives on artistic creation hinge on the assumption that art is an extension of the human experience. At its core, human-authored poetry often serves as a reflection of individual memories, traumas and joys, anchored deeply in personal history. Such poetry isn’t just a creative act; it’s a revelatory one, exposing the nuanced layers of the human psyche and allowing others to resonate with the emotional complexity presented therein (Mazza, 2016).

Thus, AI-written poetry emerges from a vastly different paradigm. Lacking personal experiences, emotions, and consciousness, AI derives its poetic compositions from algorithms, vast databases, and meticulous analysis of existing human literature. The AI doesn’t ‘feel’ or ‘reflect’ but ‘computes’ and ‘generates’. Its poetry, while often astonishing in linguistic mastery and emotional depth, is essentially a collage of human experiences, pieced together in novel ways. The crucial question then arises: Can this assemblage, devoid of personal emotional nature, be considered truly creative? (Yan et al., 2013).

A potential counterargument is that AI’s very detachment from personal experience might grant it a unique form of creativity. Without the biases, cultural impositions, and emotional boundaries that often constrict human poets, AI has the latitude to combine words, images, and sentiments in unforeseen ways. This unpredictability in generation might be perceived as a new form of creativity, one rooted not in experience but in nearly unlimited possibility. While human poets pull from their life stories, AI pulls from a collective narrative, encompassing vast dimensions of human literature (Mazzone and Elgammal, 2019).

However, many might contend that true creativity isn’t solely about the end product but also the journey of creation. The cathartic process a human poet undergoes, the interplay of memory and imagination, and the vulnerability in self-expression — all contribute to the ‘soul’ of a poem. AI, with its computational origin, lacks this journey. The absence of an emotional odyssey in the creative process might render AI-authored poetry as an impressive technical act but not a creative masterpiece in the traditional sense (Kirkpatrick, 2023).

Furthermore, the essence of creativity, as traditionally understood, entails a novelty born from introspection and external stimuli. It’s a fusion of past experiences with present perceptions, leading to unprecedented creations. AI, while unparalleled in its ability to combine and generate based on its database, might be perceived as ‘recreating’ rather than ‘creating’. It lacks the possibility of a stray thought, the weight of a haunting memory, or the spark of a sudden realization, all of which are integral to human creativity (Amabile, 2020).

In sum, the debate on whether AI-written poetry can be deemed truly creative is a complex interplay of definitions, perceptions, and values. While AI undeniably pushes the boundaries of linguistic construction and introduces novel poetic combinations, it does so without the emotional underpinning that characterizes human creativity. Whether this computational ability can be equated with human-inspired creativity remains a pivotal question in our evolving understanding of art, technology, and human experience.

Discussion

The realm of creativity, with its deep-rooted connections to human experiences, has always been multifaceted and intricate. Historically, in fields as diverse as art, science, and philosophy, creativity has been perceived as a quintessentially human attribute, a fusion of learned experiences, sudden insights, and nurturing environments. However, as we step into the era where Artificial Intelligence (AI) intersects with this deeply human trait, our perceptions and definitions of creativity are being challenged.

Artificial Intelligence, in its essence, is about processing extensive data, identifying patterns, and generating outputs based on algorithms. As AI begins to navigate the domain of poetry, it brings forth compelling questions about its genuine capacity for creativity. Can a system, built to adhere to logic, truly replicate the spontaneity and depth found in human expression? Interestingly, while AI has showcased its ability to produce poetry that can stir human emotions, the authenticity of such creations is debatable.

This debate is rooted in the very essence of creativity. If genuine creativity is the culmination of personal experiences, emotions, and unique interpretations of the world, can poetry generated by an emotionless machine be considered authentic? Moreover, the knowledge that a poem is AI-generated often influences the reader’s perception, introducing a bias that can either amplify or diminish the emotional resonance of the piece. This bias poses a relevant question: Are the emotions elicited genuine, or are they affected by our knowledge of the poem’s origin? The convergence of AI, poetry, and creativity prompts a fruitful discussion, challenging traditional perspectives and laying a path for the future of artistic expression.

Conclusion

The realm of poetic creativity, traditionally considered the pinnacle of human expressiveness, is undergoing a significant shift with the entry of AI-driven compositions. As these AI-generated poems begin to demonstrate a potential to evoke comparable emotional responses in readers, the foundation of what we consider to be ‘genuine’ or ‘authentic’ artistry is questioned. However, it is essential to note the intrinsic biases that emerge among readers when they become aware that the verses they’re engaging with are birthed from algorithms rather than human experience. As artificial intelligence continues to become an integral component of our daily lives and routines, we stand at a crossroads. We must decide how we’ll navigate this evolving landscape where technology and human creativity intersect, ensuring that we harness the strengths of both while being keenly aware of their individual limitations.

References

Aggarwal, K., Mijwil, M. M., Al-Mistarehi, A. H., Alomari, S., Gök, M., Alaabdin, A. M. Z., and Abdulrhman, S. H. (2022). Has the future started? The current growth of artificial intelligence, machine learning, and deep learning. Iraqi Journal for Computer Science and Mathematics, 3(1), pp. 115-123.

Amabile, T. M. (2020). Creativity, artificial intelligence, and a world of surprises. Academy of Management Discoveries, 6(3), pp. 351-354.

An, D., Song, Y., and Carr, M. (2016). A comparison of two models of creativity: Divergent thinking and creative expert performance. Personality and Individual Differences, 90, pp. 78-84.

Chatterjee, A., and Vartanian, O. (2016). Neuroscience of aesthetics. Annals of the New York Academy of Sciences, 1369(1), pp. 172-194.

Csikszentmihalyi, M. (2015). The systems model of creativity: The collected works of Mihaly Csikszentmihalyi. Springer.

Demmer, T.R., Kühnapfel, C., Fingerhut, J. and Pelowski, M. (2023). Does an emotional connection to art really require a human artist? Emotion and intentionality responses to AI-versus human-created art and impact on aesthetic experience. Computers in Human Behavior, pp. 1-19.

Ekman, P.E. and Davidson, R.J. (1994). The nature of emotion: Fundamental questions. Oxford University Press.

Hilse, C., Griffiths, S. and Corr, S. (2007). The impact of participating in a poetry workshop. British Journal of Occupational Therapy, 70(10), pp.431-438.

Johnson-Laird, P.N. and Oatley, K. (2022). How poetry evokes emotions. Acta Psychologica, 224, pp. 1-12.

Klausen, S.H. (2010). The notion of creativity revisited: A philosophical perspective on creativity research. Creativity research journal, 22(4), pp. 347-360.

Kirkpatrick, K. (2023). Can AI Demonstrate Creativity?. Communications of the ACM, 66(2), pp. 21-23.

Kirmani, A. R. (2022). Artificial intelligence-enabled science poetry. ACS Energy Letters, 8(1), pp. 574-576.

Köbis, N., and Mossink, L. D. (2021). Artificial intelligence versus Maya Angelou: Experimental evidence that people cannot differentiate AI-generated from human-written poetry. Computers in human behavior, 114, pp. 1 -13.

Lund, B. D., & Wang, T. (2023). Chatting about ChatGPT: how may AI and GPT impact academia and libraries?. Library Hi Tech News, 40(3), pp. 26-29.

Mazza, N. (2016). Poetry therapy: Theory and practice. Routledge.

Mazzone, M., and Elgammal, A. (2019). Art, creativity, and the potential of artificial intelligence. Arts, 8(1), pp. 2-9.

Obermeier, C., Menninghaus, W., Von Koppenfels, M., Raettig, T., Schmidt-Kassow, M., Otterbein, S., and Kotz, S. A. (2013). Aesthetic and emotional effects of meter and rhyme in poetry. Frontiers in psychology, 4, pp. 1-10.

O’Leary, D. E. (2013). Artificial intelligence and big data. IEEE intelligent systems, 28(2), pp. 96-99.

Oksanen, A., Cvetkovic, A., Akin, N., Latikka, R., Bergdahl, J., Chen, Y. and Savela, N. (2023). Artificial intelligence in fine arts: A systematic review of empirical research. Computers in Human Behavior: Artificial Humans, pp. 1-14.

Pițur, S. and Miu, A.C. (2022). Poetry-elicited emotions: Reading experience and psychological mechanisms. Psychology of Aesthetics, Creativity, and the Arts, 5(8), pp. 44-65.

Reckwitz, A. (2018). The invention of creativity: Modern society and the culture of the new. John Wiley & Sons.

Strehovec, J. (2023). From Poetry As the Excess of Language to Poems Generated by Artificial Intelligence (Poet As Researcher in the World of New Cultural Paradigms). Canadian Journal of Language and Literature Studies, 3(4), pp. 98-111.

Yan, R., Jiang, H., Lapata, M., Lin, S.D., Lv, X. and Li, X. (2013, June). I, poet: automatic chinese poetry composition through a generative summarization framework under constrained optimization. In Twenty-Third International Joint Conference on Artificial Intelligence.

Wu, Y., Mou, Y., Li, Z., and Xu, K. (2020). Investigating American and Chinese Subjects’ explicit and implicit perceptions of AI-Generated artistic work. Computers in human behavior, 104, pp.106-186.

Zhang, B., Zhu, J. and Su, H. (2023). Toward the third generation artificial intelligence. Science China Information Sciences, 66(2), pp.121-135.

About the author

Ekaterina Pretsch, MS, is a graduate of the Blekinge Technical University in Karlskrona, Sweden. She received her Master’s degree in Strategic Leadership towards Sustainability in 2015. Since then, she has been working on the topic of sustainable development both in academia and corporate sector. Currently, she works on research projects about artificial intelligence at the Austrian-based research firm, ERC Experience Research and Consulting. Her academic and professional pursuits revolve around the topics of sustainable development, artificial intelligence, creativity and machine learning.