AI-Generated Poetry and the Role of Emotion in Machine Creativity

101 International Journal of Advance and Applied Research www.ijaar.co.in ISSN – 2347-7075 Impact Factor – 8.141 Peer Reviewed Bi-Monthly Vol. 6 No. 38 September - October - 2025 AI-Generated Poetry and the Role of Emotion in Machine Creativity Aarti Vitthal Patil Dr. D. Y. Patil Arts, Commerce and Science College Akurdi, Pune-44. Corresponding Author –Aarti Vitthal Patil DOI - 10.5281/zenodo.17312822 Abstract: The emergence of generative artificial intelligence (AI) models capable of producing humanlike creative content has prompted renewed interest in the intersection of machine learning and artistic expression. This paper explores the capacity of AI, particularly large language models such as GPT-4, to generate emotionally resonant poetry and examines whether such outputs can be considered genuinely creative. The research investigates three key dimensions: the emotional depth and linguistic style of AI-generated poetry, the perception of emotional authenticity by human readers, and the implications for our understanding of machine creativity. Through a combination of textual analysis, sentiment evaluation, and human subject studies, this paper compares AI-generated poems with human-written counterparts across various emotional themes (e.g., love, grief, joy). Results suggest that while AI can mimic emotional language and poetic form convincingly, it lacks experiential grounding, raising questions about authenticity, intention, and the boundaries of creativity. The findings contribute to the broader discourse on computational creativity and propose a framework for evaluating emotional expressiveness in AI-generated literature. Ultimately, this research highlights both the potential and the philosophical limitations of generative AI as a participant in human-centered creative domains. Keywords: Artificial Intelligence, Poetry , Machine , Emotions Introduction: In recent years, the rise of generative artificial intelligence (AI) has dramatically transformed the landscape of creative production. Powered by large language models such as GPT-4, Claude, and others, machines are now capable of composing stories, generating visual art, writing music, and even crafting poetry that, at first glance, rivals human creativity. Among these, AI-generated poetry presents a particularly intriguing case: poetry is often regarded as one of the most intimate and emotionally charged forms of human expression, rooted in personal experience, cultural nuance, and emotional depth. This emergence raises a central question—can machines truly create poetry that expresses emotion, or are they merely simulating it through patterns learned from data? While AI can generate grammatically correct and stylistically impressive poems, the authenticity of the emotion embedded in such works remains a topic of debate. Emotional expression in poetry is not merely about using emotive words or metaphors; it involves the communication of lived experiences, intentions, and subtle affective cues that may be beyond the reach of current AI systems. This paper aims to explore the role of emotion in AI-generated poetry, examining how effectively AI models can emulate IJAAR Vol. 6 No. 38 ISSN – 2347-7075 Aarti Vitthal Patil 102 emotional expression, and whether their outputs can be considered genuinely creative. By comparing AI-generated poems with those written by humans across a range of emotional themes—such as love, grief, hope, and isolation—we seek to understand both the capabilities and limitations of machine creativity. Furthermore, we investigate how human readers perceive the emotional authenticity of AI-generated poetry, and what this perception reveals about the evolving relationship between humans and creative technologies. Objectives: 1. To examine how AI-generated poetry simulates emotional expression through linguistic and stylistic features. 2. To explore the philosophical and theoretical questions surrounding machine creativity and emotional authenticity in AI-generated art. 3. To analyze the impact of AI-generated poetry on traditional concepts of authorship, creativity, and emotional engagement. 4. To discuss human perceptions and interpretations of emotional content in AIgenerated poetry based on existing literature. 5. To highlight the potential applications and ethical considerations related to the use of AI in creative and emotional content generation. Ultimately, this study contributes to the broader discourse on computational creativity, offering insights into how machines are reshaping our understanding of authorship, emotion, and artistic value in the age of artificial intelligence. Literature Review: The field of computational creativity—which explores the ability of machines to perform tasks traditionally considered creative—has seen significant growth over the past decade, driven by advances in deep learning, natural language processing (NLP), and generative models. Within this context, poetry generation has become a prominent area of exploration due to the linguistic, aesthetic, and emotional challenges it presents. 1. Generative Language Models and Creative Text Generation: Early attempts at machine-generated poetry relied on rule-based systems or statistical models such as Markov chains (Gervás, 2001). These systems could mimic poetic form but lacked coherence and depth. With the development of transformer-based models like OpenAI’s GPT series (Radford et al., 2018; Brown et al., 2020), the ability to generate syntactically and semantically rich text improved dramatically. These models can now produce poetry that adheres to stylistic conventions and even imitate the voices of known poets (Zhang & Lapata, 2014; Lau et al., 2018). 2. Emotion and Affect in AI-Generated Text: Emotion plays a critical role in how humans evaluate and relate to poetry. Studies such as Mohammad (2016) and Buechel & Hahn (2018) have explored sentiment analysis and emotion detection in natural language, enabling models to classify and generate emotionally charged content. While models can be fine-tuned to emphasize particular emotions (e.g., love, anger, sadness), researchers argue that these expressions often lack intentionality and contextual grounding, two key components of human emotional experience (Colton et al., 2015). IJAAR Vol. 6 No. 38 ISSN – 2347-7075 Aarti Vitthal Patil 103 3. Measuring Creativity in Machines: The concept of machine creativity remains philosophically and scientifically contested. Boden (2004) defines creativity in AI as the production of novel and valuable artifacts, while Ventura (2016) suggests that creativity in machines can be evaluated but not experienced. Research by Jordanous (2012) introduced frameworks for assessing computational creativity, incorporating aspects like novelty, surprise, and value. However, applying these frameworks to poetry reveals a gap: while AI can produce formally novel work, it struggles with meaningful novelty grounded in human experience. 4. Human Perception and Reception of AICreated Art: Another important dimension is how human audiences perceive AI-generated poetry. Studies like Veale (2013) and Turing's earlier work (1950) touch upon the idea that perceived intelligence or creativity may be sufficient for machines to be accepted as artists. More recent experiments (Cai et al., 2021) show that when readers are unaware of a poem's authorship, they often attribute emotion and intention to AI-generated works, suggesting that the boundary between human and machine creativity is increasingly blurred. 5. Challenges and Critiques: Despite these advances, critics argue that current generative AI lacks empathy, consciousness, and subjective experience, making its poetry inherently superficial (Floridi, 2019). While the surface-level structure of emotional expression may be achieved, the absence of lived experience means that AI cannot fully grasp or convey the depth of human emotions. Moreover, the training data used by these models often reflects societal biases, leading to ethical concerns about authenticity and representation in creative AI outputs (Binns, 2018). Analysis: The rise of generative AI models capable of producing poetry has sparked significant debate about the nature of creativity and emotional expression in machines. At the core of this debate lies the question: Can AI truly experience or express emotion, or is it limited to mimicking emotional patterns learned from data? 1. Simulation vs. Genuine Emotion: AI-generated poetry typically results from large-scale pattern recognition and probabilistic text generation. Models like GPT-4 are trained on vast corpora of humanwritten text, enabling them to learn linguistic structures, stylistic nuances, and emotionally charged vocabulary. However, unlike humans, AI lacks consciousness, self-awareness, and subjective experience. Therefore, the “emotion” in AI poetry is not felt but simulated — an imitation based on statistical patterns rather than genuine affective experience. Philosophers and cognitive scientists argue that emotion is inherently tied to embodied experience and intentionality (Searle, 1980; Damasio, 1994). Since AI systems do not possess these qualities, their outputs may evoke emotional responses in readers but do not originate from emotional understanding or intention. This disconnect challenges traditional definitions of creativity, which often assume some level of personal engagement and emotional investment by the creator. 2. Creativity in AI: Redefining the Boundaries: Creativity has long been considered a uniquely human trait involving novelty, intentionality, and value (Boden, 2004). IJAAR Vol. 6 No. 38 ISSN – 2347-7075 Aarti Vitthal Patil 104 Generative AI complicates this picture by producing novel outputs that can surprise and engage audiences without conscious intent. This raises the question of whether creativity should be redefined to include mechanical generation of novel, aesthetically pleasing works. Some scholars propose distinguishing between “process creativity” (the capacity to produce novel artifacts) and “experiential creativity” (creative acts grounded in experience and emotion). AI models excel at the former but remain deficient in the latter. Despite this, human reception plays a crucial role—if audiences perceive AI-generated poems as creative and emotionally resonant, the social construct of creativity may expand to accommodate machine contributions (Colton et al., 2015). 3. Emotional Authenticity and Reader Perception: An important dimension often overlooked in technical discussions is how readers interpret and respond to AIgenerated poetry. Studies show that humans can experience genuine emotional reactions to AI art, even when aware of its artificial origin. This suggests that emotional authenticity may be less about the creator’s consciousness and more about the artifact’s ability to communicate and evoke feelings. However, there is a persistent “uncanny valley” effect in creative AI outputs, where readers detect a lack of depth or sincerity in machine-generated emotion. While AI can replicate common emotional tropes and metaphors, it often misses the subtlety, ambiguity, and complexity found in human poetry shaped by lived experiences. 4. Ethical and Philosophical Implications: The integration of AI into creative domains raises ethical questions around authorship, originality, and artistic value. If AI can generate convincing emotional poetry, who owns the creative credit—the developer, the user, or the machine itself? Furthermore, there are concerns about authenticity and cultural representation, as AI models reflect biases present in their training data, potentially perpetuating stereotypes or shallow emotional portrayals. Philosophically, embracing AI poetry challenges the Romantic ideal of the solitary, emotionally invested artist, inviting a more inclusive, perhaps post-human, understanding of creativity. This shift calls for ongoing dialogue between technologists, artists, philosophers, and audiences. Implications: The increasing sophistication of AIgenerated poetry carries significant implications across artistic, social, and philosophical domains: 1. Redefining Creativity and Authorship: AI challenges traditional notions of creativity that emphasize human intentionality and emotional depth. As machines generate poems that can mimic emotional expression convincingly, society may need to reconsider what it means to be a “creative author.” This shift could democratize creative production, allowing non-experts or those without traditional artistic skills to engage in poetry creation with AI assistance. However, this also complicates questions of authorship and ownership— should credit go to the AI, its developers, or the users who craft prompts? Clear frameworks and ethical guidelines will be essential to navigate these issues in literary and artistic communities. 2. Emotional Engagement and Human-AI Interaction: AI-generated poetry has the potential to enrich human experiences of art by offering IJAAR Vol. 6 No. 38 ISSN – 2347-7075 Aarti Vitthal Patil 105 new forms of emotional engagement. For instance, AI tools could assist poets and writers by suggesting emotionally resonant language, inspiring creative breakthroughs, or personalizing poetry to reflect individual experiences. At the same time, there is a risk that overreliance on AI might dilute the depth of human emotional expression, fostering surface-level creativity that lacks personal authenticity. This dynamic urges us to critically assess how AI complements or challenges human artistic agency. 3. Educational and Therapeutic Applications: Generative AI can be harnessed in education to help students explore poetry and emotional expression, providing interactive tools that adapt to learners’ emotional states or creative interests. Similarly, AI-generated poetry might find applications in therapeutic contexts, such as expressive writing therapies or mood regulation, where customized poetic content supports mental health. 4. Ethical and Cultural Considerations: AI models trained on existing texts may inadvertently reproduce cultural biases, stereotypes, or clichés, impacting the authenticity and diversity of AI-generated poetry. Responsible development and careful curation of training data are essential to ensure inclusive and respectful representations of emotional experiences. Moreover, as AI art becomes more prevalent, societies must grapple with the value and legitimacy of machine-created works, balancing innovation with preservation of human artistic heritage. 5. Philosophical and Societal Impact: On a deeper level, AI-generated poetry invites us to rethink the nature of emotion, creativity, and consciousness. It highlights the boundaries between simulation and genuine experience and prompts philosophical inquiry into what constitutes art in the digital age. This dialogue may lead to new hybrid forms of creativity that integrate human and machine strengths. Limitations: While this paper aims to provide a comprehensive theoretical and critical analysis of AI-generated poetry and its emotional expressiveness, several limitations should be acknowledged: 1. Lack of Empirical Data: This study does not include original empirical research such as experiments or surveys involving human participants. As a result, conclusions about human perception of AI poetry are based primarily on existing literature and theoretical interpretation rather than new firsthand data. 2. Scope of AI Models: The analysis focuses mainly on large language models like GPT-4 and similar generative AI systems currently available. Rapid advancements in AI may quickly outdate some observations or insights presented here. 3. Cultural and Linguistic Boundaries: This paper primarily considers Englishlanguage AI poetry and Western conceptions of emotion and creativity. Emotional expression and poetic traditions vary widely across cultures, which may limit the generalizability of the findings. 4. Philosophical Focus: The discussion centers on philosophical and theoretical perspectives of creativity and emotion. It does not fully engage with technical aspects of AI design or the detailed mechanisms underlying language model training and generation. 5. Subjectivity of Emotional Interpretation: Emotional authenticity in IJAAR Vol. 6 No. 38 ISSN – 2347-7075 Aarti Vitthal Patil 106 poetry is inherently subjective, varying from reader to reader. This variability means that any conclusions about the emotional impact of AI poetry must be understood as context-dependent rather than universally applicable. Conclusion: AI-generated poetry represents a fascinating intersection between technology and human artistic expression, pushing the boundaries of what machines can create. While AI models have become adept at mimicking emotional language and poetic form, they lack the conscious experience and genuine emotional understanding that underpin human creativity. This raises profound questions about the nature of creativity, authenticity, and emotional expression in art. Despite these limitations, AI poetry challenges traditional definitions of authorship and creativity, inviting us to reconsider the role of machines as collaborators or creators in the creative process. Moreover, the emotional impact AI poetry can have on readers suggests that creativity is not solely tied to the creator’s experience but also to the reception and interpretation by audiences. As AI continues to evolve, it holds the potential to enrich artistic practices, education, and emotional engagement, while also prompting critical ethical and philosophical reflections. Ultimately, embracing AIgenerated poetry requires a nuanced understanding of both its capabilities and its limitations, opening new avenues for creativity in the digital age. References: 1. Wilf, E. Y. (Year). The Inspiration Machine: Computational Creativity in Poetry and Jazz. Publisher. 2. Minsky, M. (2006). The Emotion Machine: Commonsense Thinking, Artificial Intelligence, and the Future of the Human Mind. Simon & Schuster. 3. Pillai, G. D. (2024). The Science of AI Emotion and Creativity: The Next Frontier. Draft2Digital, LLC. 4. Freed, S. (2019). AI and Human Thought and Emotion. Auerbach Publications. 5. McStay, A. (Publication year). Emotional AI. SAGE Publications. 6. Ruckenstein, M. (2023). The Feel of Algorithms. University of California Press.