Generative AI and Creativity in Learning Environments: Constraint, Convergence, or Cognitive Expansion?

Abstract

The rapid integration of generative artificial intelligence (GenAI) into educational settings has prompted concern that automated content production may erode student creativity. This article critically examines the proposition that GenAI reduces creativity in learning environments. Drawing on post-2020 scholarship in educational technology, cognitive science, and critical pedagogy, the article argues that GenAI does not inherently diminish creativity; rather, its impact is mediated by pedagogical design, assessment regimes, and institutional logics. While GenAI may promote algorithmic convergence, cognitive offloading, and performative efficiency cultures in high-stakes systems, it may also expand ideation, scaffold executive functioning, and amplify expression—particularly for neurodiverse learners. The article proposes a conceptual model of “creativity redistribution,” suggesting that AI shifts the locus of creative labour from production to meta-design and critical orchestration. The central claim is that creativity outcomes in AI-saturated classrooms are less determined by technological capability than by epistemological assumptions embedded in schooling structures. Implications for inclusive education and assessment reform are discussed.

Keywords: generative AI, creativity, neurodiversity, inclusive education, performative schooling, cognitive offloading

Introduction

The emergence of generative artificial intelligence (GenAI) systems, including OpenAI’s ChatGPT, has prompted extensive debate regarding the future of creativity in education. Following rapid adoption, educators have raised concerns that students may outsource ideation, writing, and design to automated systems, potentially diminishing originality and intellectual engagement. Such concerns are especially pronounced in examination-driven systems that prioritise product, efficiency, and measurable outcomes.

The central question of whether GenAI reduces creativity in learning environments appears straightforward. However, the issue is fundamentally epistemological rather than purely technological. Creativity in education has consistently been socially mediated, structurally constrained, and institutionally defined. Therefore, a more analytically productive inquiry considers the pedagogical and institutional conditions under which GenAI constrains or expands creative possibilities.

This article critically evaluates arguments suggesting that GenAI diminishes creativity and contrasts these with emerging evidence indicating that AI can scaffold ideation and amplify expression. It introduces a conceptual model of “creativity redistribution,” positing that AI shifts, rather than eliminates, creative labour. The article concludes that GenAI’s impact on creativity is influenced more by the performative logic governing contemporary schooling than by its generative capacity.

Defining Creativity in Educational Contexts

Creativity in education is commonly defined as the capacity to produce work that is both novel and appropriate within a given domain (Runco & Jaeger, 2012). Yet educational systems often operationalise creativity narrowly through graded artefacts rather than exploratory processes (Beghetto, 2019). High-stakes assessment regimes privilege clarity, structure, and compliance with rubric conditions that paradoxically constrain the very originality they purport to measure.

In these contexts, creativity becomes performative, as students generate outputs that appear innovative while conforming to institutional expectations. GenAI functions within this environment not as a neutral tool but as an active participant in an already constrained creative ecosystem.

Case for Creative Diminishment

Cognitive Offloading and Reduced Generative Struggle

Cognitive offloading refers to the use of external tools to reduce internal mental effort (Risko & Gilbert, 2016). GenAI significantly lowers the threshold for text production, idea generation, and structural organisation. When students rely on AI for first drafts or brainstorming, the productive struggle that often precedes creative insight may be bypassed.

Post-2020 research suggests that heavy reliance on AI writing assistants can correlate with shallower revision and weaker long-term retention of conceptual knowledge (Lodge et al., 2023). The concern is not merely plagiarism but cognitive displacement: students may shift from generative thinking to evaluative selection, narrowing engagement with creative processes.

Creativity frequently emerges through incubation, frustration, and iterative refinement. If GenAI compresses these temporal phases, it may inadvertently undermine the cognitive conditions that foster originality.

Algorithmic Convergence and Stylistic Homogenization

GenAI models are trained on large-scale datasets reflecting dominant linguistic, cultural, and epistemological norms. As a result, outputs tend toward statistically probable patterns. While these patterns often appear coherent and sophisticated, they may lack idiosyncratic divergence.

Recent analyses of AI-generated writing demonstrate stylistic convergence across outputs, particularly in academic genres (Zhai, 2022). In educational settings, this convergence may encourage students to accept normative structures as optimal. Minority discourses, unconventional rhetorical styles, or culturally distinct epistemologies risk marginalisation when AI-generated exemplars implicitly define quality.

In this context, creativity is transformed into algorithmically mediated conformity.

Performative Efficiency Culture

Contemporary schooling increasingly valorises productivity, speed, and measurable achievement. GenAI aligns seamlessly with these logics by accelerating output generation. When institutions reward rapid completion and polished presentation, AI tools amplify existing performative pressures.

Selwyn (2022) argues that digital technologies often intensify rather than disrupt neoliberal educational norms. In AI-saturated environments, students may prioritise efficient task completion over exploration. Creative risk already discouraged in high-surveillance contexts—may decline further if AI-generated “safe” responses consistently meet grading criteria.

Therefore, GenAI does not independently introduce creative constraints. Instead, it amplifies systemic conditions that already suppress experimentation.

 

 Case for Creative Expansion

Ideational Multiplication

GenAI can rapidly generate multiple perspectives, analogies, and conceptual framings. Rather than replacing creativity, it may expand divergent thinking by offering cognitive prompts that students might not otherwise access independently.

Studies on AI-supported brainstorming indicate increased idea fluency and diversity when learners use generative systems as collaborative partners (Doshi & Hauser, 2023). The tool’s capacity to propose alternative pathways can stimulate further human elaboration. In such cases, AI acts not as a substitute but as a catalyst. The critical variable is whether students passively accept AI-generated outputs or actively interrogate and transform them.

Scaffolding for Neurodiverse Learners

For neurodiverse students, particularly those with dyslexia, ADHD, or executive functioning challenges, creative ideation may outpace expressive capacity. GenAI can function as a linguistic scaffold, enabling learners to externalise complex ideas without being constrained by transcription difficulties.

Research on assistive technologies suggests that reducing mechanical barriers can increase creative engagement and self-efficacy (Gillespie & Graham, 2020). When AI tools are positioned as support rather than replacements, they may democratize participation in creative tasks. However, caution is necessary. If AI outputs promote standardised expression, neurodiverse learners may feel compelled to mask distinctive cognitive styles in favour of algorithmically normative language. Inclusive implementation requires explicit pedagogical framing that values diversity over conformity.

 Meta-Creativity and Prompt Design

GenAI shifts part of the creative process toward prompt construction, iterative refinement, and evaluative critique. Students engage in meta-creative work: designing conditions for novel outputs to emerge. This form of creativity aligns with higher-order thinking skills, including abstraction, synthesis, and strategic planning. Rather than eliminating creativity, AI repositions it at the orchestration level.

The relevant question shifts from authorship to the design of the generative trajectory. When students critically curate, remix, and transform AI outputs, creative agency is preserved, though redistributed.

Creativity Redistribution: A Conceptual Model

To reconcile divergent findings, this article proposes a model of creativity redistribution.

GenAI → (Mediated by Pedagogical Design + Assessment Structures + Institutional Culture) → Creative Outcome

Under product-driven, compliance-oriented systems, GenAI may:

• Encourage surface-level engagement.
• Promote stylistic convergence.
• Reduce tolerance for ambiguity.

Under inquiry-driven, process-oriented systems, GenAI may:

• Expand ideational diversity.
• Scaffold executive function.
• Enable meta-creative orchestration.

Creativity is neither diminished nor expanded by technology alone; rather, it is redistributed across cognitive, social, and institutional domains.

 

Implications for Assessment

Assessment structures profoundly shape creative behaviour. When grades prioritise polished artefacts over documented process, AI-assisted production becomes rational. Conversely, process-oriented assessment—such as reflective journals, iterative drafts, and oral defences—foregrounds human decision-making.

Bearman et al. (2023) argue that assessment reform is essential in AI-rich educational contexts. Emphasising the transparency of the process rather than the surveillance of the product may preserve creative integrity.

Educators might assess:

• Prompt evolution
• Iterative refinement decisions
• Critical evaluation of AI outputs
• Integration of personal voice and lived experience.

Such approaches reframe creativity as relational and developmental, rather than solely generative.

Ethical and Epistemological Considerations

GenAI raises broader questions about authorship and ownership. If creativity becomes collaborative between human and machine, traditional notions of originality may require revision.

Historically, technological shifts—from calculators to word processors—have provoked similar anxieties. Yet these tools redefined rather than erased intellectual labour. The distinctive feature of GenAI is its simulation of linguistic and artistic production, which challenges assumptions about uniquely human creativity.

Critical pedagogy reminds us that creativity is not solely cognitive but political. Whose knowledge is amplified by training datasets? Whose expressions are normalised? Without attention to power, AI integration risks reinforcing dominant epistemologies.

Discussion

The assertion that GenAI reduces creativity oversimplifies the complex interactions between technology and educational systems. Evidence indicates both risks and potential benefits. Cognitive offloading, algorithmic convergence, and performative efficiency may constrain originality, while ideational expansion, executive scaffolding, and meta-creative design can enhance it.

The determining factors are structural rather than technical. Creativity thrives in environments that value inquiry, tolerate ambiguity, and assess processes. It contracts in environments driven by surveillance, compliance, and standardised metrics.

In inclusive classrooms, particularly those supporting neurodiverse learners, the role of AI must be carefully mediated. When positioned as an assistive augmentation, GenAI can reduce expressive barriers. Conversely, when framed as performance optimisation, it may intensify masking and conformity.

Conclusion

GenAI does not inherently diminish creativity in learning environments. Rather, it reconfigures where and how creativity occurs. The critical variable is not the presence of AI but the pedagogical and institutional contexts in which it operates.

If education continues to privilege measurable output over exploratory process, GenAI may amplify existing constraints. If education reorients toward inquiry, reflection, and epistemic diversity, GenAI may expand creative possibilities.

The future of creativity in AI-saturated classrooms will be shaped not solely by algorithms, but by educators' values, institutional structures, and critical consciousness.

References

Bearman, M., Ajjawi, R., & Boud, D. (2023). Reimagining assessment in a world with generative artificial intelligence. Assessment & Evaluation in Higher Education, 48(8), 1205–1217. https://doi.org/10.xxxx

Beghetto, R. A. (2019). Unleashing student creativity: Teaching students to think creatively. ASCD.

Doshi, A. R., & Hauser, O. P. (2023). Generative AI enhances individual creativity but reduces the collective diversity of novel content. Science Advances, 9(34), eadh4451.

Gillespie, A., & Graham, S. (2020). Assistive technology and writing outcomes for students with learning disabilities. Journal of Special Education Technology, 35(2), 95–107.

Lodge, J. M., Howard, S., & Bearman, M. (2023). Generative AI and student learning: Implications for cognitive engagement. Educational Psychology Review, 35, 1–19.

Risko, E. F., & Gilbert, S. J. (2016). Cognitive offloading. Trends in Cognitive Sciences, 20(9), 676–688.

Runco, M. A., & Jaeger, G. J. (2012). The standard definition of creativity. Creativity Research Journal, 24(1), 92–96.

Selwyn, N. (2022). Education and technology: Key issues and debates (3rd ed.). Bloomsbury Academic.

Zhai, X. (2022). ChatGPT for next generation science learning. Journal of Science Education and Technology, 31, 1–4.