Can artificial intelligence be creative?

May 17, 2026··

17 min read

Can Artificial Intelligence Be Creative?

Executive summary
Artificial intelligence (AI) systems today can and do produce outputs that many observers deem creative: paintings, music, poems, product designs, scientific hypotheses, and even humorous quips. Whether AI is truly "creative" depends on how one defines creativity. If creativity is taken to mean the generation of novel, surprising, and valuable artifacts or ideas according to social standards, then contemporary AI systems exhibit many hallmarks of creativity, especially as collaborators and amplifiers of human imagination. If creativity requires consciousness, intentionality, subjective experience, or moral agency, then present-day AI does not meet that bar. The practical and theoretical landscape is rich and rapidly evolving: computational methods (GANs, transformers, evolutionary algorithms), evaluation metrics, legal frameworks, and societal impacts are all in flux. This article provides an in-depth, multidisciplinary examination of the question—covering history, theory, techniques, empirical results, limitations, legal/ethical issues, and future prospects—so researchers, practitioners, policymakers, and interested readers can form an informed view.

Contents

Definitions and frameworks for creativity
Historical background: creativity and computing
Theoretical foundations and models
Computational approaches and algorithms
Applications and examples across domains
Evaluation: how to measure machine creativity
Philosophical objections and conceptual limits
Legal, ethical, and social implications
Current state-of-the-art and notable systems
Future trajectories and research directions
Practical recommendations and concluding remarks
Further reading

1. Definitions and frameworks for creativity

Why definitions matter

The answer to "Can AI be creative?" is definition-sensitive. Different fields emphasize different components: novelty, value, surprise, process, intentionality, or social recognition.
Clear operational definitions are crucial for research design, evaluation, legal assessment, and public communication.

Common conceptualizations

Novelty + Value: The canonical pragmatic definition from psychology and computational creativity: a product is creative if it is novel (or unexpected) and valuable (useful, aesthetically pleasing, or meaningful) in a given context (Sternberg & Lubart; Rhodes).
Boden's taxonomy (Margaret Boden, 1990s–2000s):
- Combinational creativity: new combinations of familiar ideas.
- Exploratory creativity: exploring structured conceptual spaces to find new possibilities.
- Transformational creativity: changing the rules or conceptual space itself to enable new kinds of artifacts.
Systems view (Mihaly Csikszentmihalyi): Creativity arises from the interaction of three elements—individual, domain (symbolic rules and materials), and field (gatekeepers and social recognition). This emphasizes social validation as key.
Process-oriented vs Product-oriented:
- Process-oriented: Focus on cognitive mechanisms (divergent thinking, associative processes).
- Product-oriented: Focus on outcome characteristics (novelty, appropriateness).
Intentionality and phenomenology: Philosophers emphasize subjective experience, intention, and agency—dimensions in which machines arguably lack status.

Operational metrics frequently used:

Novelty, surprise, divergence, fluency, flexibility, and usefulness.
Psychometric tests adapted from human creativity: Torrance Tests of Creative Thinking (TTCT), Remote Associates Test (RAT), etc., sometimes used as inspiration for evaluating AI.

2. Historical background: creativity and computing

Early optimism and symbolic AI (1950s–1980s)

Early AI pioneers (e.g., Newell & Simon) tackled problems like theorem proving and story generation, suggesting rule-based systems could emulate creative problem solving.
Programs like ELIZA (Weizenbaum, 1966) and story/planning systems (e.g., TALE-SPIN) revealed that simple rule-based approaches could produce outputs perceived as humanlike in limited settings—sparking debate about simulation versus genuine creativity.

Computational creativity as a field (1990s–present)

Emergence of "computational creativity" as an explicit subfield with conferences (ICCC), journals, and frameworks (e.g., Wiggins' search/transformational frameworks).
Margaret Boden’s work (The Creative Mind, 1990; 2004) synthesized conceptual distinctions used throughout the discipline.

Statistical learning and generative models (2010s–2020s)

The deep learning revolution changed the toolkit. Neural networks, especially generative models like GANs (Goodfellow et al., 2014), VAEs (Kingma & Welling, 2013), and Transformers (Vaswani et al., 2017), allowed systems to produce high-quality images, audio, and text.
Systems like DeepDream (2015), neural style transfer (Gatys et al., 2015), and large language models (GPT series) brought AI-produced art and writing into public awareness.
Legal and economic debates intensified as AI-generated images and texts entered commercial markets (e.g., lawsuits over dataset scraping 2023–2024).

3. Theoretical foundations and models

Computational models of creative cognition

Search processes: Creativity as search in large, structured spaces (Boden; Newell & Simon). Heuristics guide search to regions of novelty.
Heuristic combinatorics and association: Creativity as recombination guided by probabilistic associations (semantic networks, vector spaces).
Evolutionary computation: Variation + selection processes (genetic algorithms, genetic programming) simulate aspects of biological creativity.
Exploration–exploitation trade-offs: Reinforcement learning and bandit frameworks capture the tension between trying familiar options and exploring novel ones.
Bayesian and probabilistic models: Model how agents infer surprising hypotheses, make analogies, or generate counterfactuals.

Cognitive and neurocognitive perspectives

Divergent thinking, associative networks, and cognitive control interplay in human creativity. Some models propose a dual-process framework: spontaneous associative processes (idea generation) + controlled refinement (evaluation and selection).
AI analogues: generative coupling (creative ideation) combined with discriminative evaluation modules (critics, classifiers, human feedback).

Socio-cultural models

Creativity is validated by the field: the same artifact might be creative in one domain and ordinary in another. AI's outputs acquire creative status when recognized by audiences, critics, or markets.

Philosophical stances

Strong AI creativity: Machines are or can become creative agents.
Weak AI creativity: Machines simulate creativity; genuine creativity requires intentional consciousness, which machines lack.
Anthropocentric and functionalist positions: If behavior is functionally equivalent, one might ascribe creativity regardless of subjective states.

4. Computational approaches and algorithms

Overview

Contemporary creative AI systems typically combine generative modules (to create possibilities) and evaluative modules (to judge, filter, or refine). They may use search, learning, or hybrid methods.

Key algorithmic families

Generative Adversarial Networks (GANs)

Two networks: a generator creates samples; a discriminator distinguishes real from generated.
Strengths: high-fidelity, realistic images; latent space interpolation enables exploration.
Examples: StyleGAN, BigGAN. Used in visual art and design.

Variational Autoencoders (VAEs) and Latent Variable Models

Model data via continuous latent codes; support controlled manipulation and interpolation.
Useful for smooth exploration of style/structure.

Diffusion models

Sequential denoising models (e.g., Stable Diffusion) that produce high-quality images from noise conditioned on text.
Excellent fidelity and controllable via conditioning and text prompts.

Transformers and Large Language Models (LLMs)

Attention-based sequence models that learn statistical regularities over massive text corpora.
Able to generate coherent text, poetry, dialogues, and code; used for idea generation and creative writing.

Symbolic and rule-based systems

Capture domain knowledge, constraints, and grammar—useful for structured creative tasks (e.g., music composition with counterpoint rules).

Evolutionary and population-based methods

Generate diverse candidate artifacts via mutation and recombination; evaluate via fitness functions (aesthetic metrics, human feedback).

Reinforcement learning and generative RL

Agents generate sequences optimized for some reward (e.g., novelty score, human preference).

Hybrid systems

Combine symbolic planning, machine learning, and human-in-the-loop evaluation for stronger performance and controllability.

Simple pseudocode: creative search with evolutionary algorithm

Plain Text

Initialize population P with random candidate artifacts
for generation in 1..N:
  Evaluate fitness for each candidate (e.g., novelty + aesthetics + constraints)
  Select parents proportionally to fitness
  Recombine and mutate to produce offspring
  Possibly inject random new individuals or human-guided seeds
  Replace least fit individuals with offspring
Return best candidates

Prompting and conditioning

Text and multimodal conditioning (prompts, style tokens, control images) are practical tools for steering creative production in models like GPT and diffusion models.

Interactive/co-creative architectures

Systems designed for collaboration: the AI proposes variants, the human selects or refines; the loop iterates. Examples: AI-assisted music composition tools, sketch-to-design systems.

5. Applications and examples across domains

Visual arts and design

Neural style transfer, DeepDream, GAN art, diffusion-model art (DALL·E, Midjourney, Stable Diffusion).
Examples: AI-generated portraits sold in galleries and auctions; architects use generative models to explore forms.

Music and audio

Systems: Magenta (Google), AIVA, OpenAI Jukebox, RNN and Transformer-based music generation.
Use cases: composing melodies, accompaniment, sound design, adaptive game music.

Literature and writing

LLMs (GPT-3/4/4o, Claude, Llama) generate short stories, poems, scripts, and marketing copy.
Tools for brainstorming plot points, rewriting content, or ideation.

Film and animation

Storyboarding assistance, shot planning, script suggestions, and automated animation from text prompts.

Scientific discovery and research

Generative models and search algorithms propose hypotheses, chemical compounds, or protein structures (e.g., AlphaFold's transformative prediction of protein folding—its outputs can accelerate novel scientific work, though whether this is "creative" is debated).

Product design and engineering

Generative design tools (Topology optimization, Autodesk's generative design) produce novel structures optimized for constraints like weight and strength.

Video games and interactive media

Procedural content generation (PCG) for levels, narratives, characters. AI systems explore vast spaces of game artifacts.

Education and creativity-enhancing tools

AI tutors and ideation systems that prompt divergent thinking and scaffold creative exercises.

Examples illustrating different forms of creativity

Combinational: Remixing styles (Van Gogh style applied to contemporary photos).
Exploratory: Traversing latent space to discover new visual motifs.
Transformational: Generating an art form that alters its own rules—a rarer and higher bar (some claim style evolution in long-running generative adversarial setups approximates this).

6. Evaluation: how to measure machine creativity

Why evaluation is hard

Creativity is context-dependent and partly socially validated.
Automated metrics may capture surface qualities but not cultural or conceptual value.

Common evaluation strategies

Human judgment

Crowdsourced ratings or expert critique on novelty, aesthetic quality, and meaningfulness.
Comparative studies showing humans rate AI outputs as creative under certain conditions.

Psychometric-inspired tests

Adapting human creativity tests (RAT, TTCT) to AI: e.g., assessing associative richness, ability to generate remote associations.

Computational proxies

Novelty scores: distance in latent/feature space from training examples.
Surprise: low predicted probability under a model.
Diversity and fluency: measures like entropy, distinct-n-gram, or Inception Score / FID for images.
Value: task-specific metrics (e.g., chemical property optimizations).

Process-oriented evaluation

Inspecting search trajectories, latent manipulations, or the inferential chain to assess whether the process exhibits generative hallmarks.

Longitudinal and cultural impact measures

Citation, adoption, recognition by experts and institutions (galleries, awards), market success.

Challenges and caveats

Novelty metrics can be gamed; high novelty without coherence is useless.
Human evaluation is expensive and subjective; cultural biases shape judgments.
Attribution of creativity in co-creative scenarios (who gets credit: human author, tool provider, dataset curators?) complicates evaluation.

Sample evaluation protocol (high level)

Define creative criteria (novelty, value, surprise) appropriate to domain.
Use a mix of automated metrics and human expert ratings.
Test against baselines (random, human-generated, simpler algorithms).
Report process transparency, dataset provenance, and human involvement.

7. Philosophical objections and conceptual limits

Key objections

The intentionality objection

Machines lack intentions, desires, or agency; they do not "intend" to create, so their outputs are not genuinely creative.

The understanding objection (Searle’s Chinese Room-style)

Systems may manipulate symbols without understanding; producing outputs that mimic human creativity is not the same as genuine understanding.

The authorship and ownership objection

If creativity requires agency tied to moral/legal responsibility, who is the creator: the human user, the model developer, or the dataset sources?

The consciousness and qualia objection

Some argue creativity is inseparable from subjective experience (aesthetic appreciation, emotion), which machines lack.

Responses and counterarguments

Functionalism: If a system produces outputs that meet the criteria (novel, valuable) and participates in social processes of validation, it can be regarded as creative regardless of internal qualia.
Gradualism: Creativity can be seen as a spectrum: AI can be more or less creative depending on capacities, even if not fully equivalent to human creative consciousness.
Sociocultural perspective: Creativity is attributed socially; if communities accept AI outputs as creative, that social attribution matters.

Hybrid views

Distinguish "creativity-as-output" (machines already excel) from "creativity-as-agentive-process" (machines currently fall short). Focus on human-AI co-creativity as a pragmatic middle path.

Legal status and IP

Copyright: Jurisdictions vary, but many legal systems link authorship to natural persons. Recent cases and policy discussions indicate AI-only works may not receive copyright; human contribution or direction is often required.
Dataset and training data issues: Scraping copyrighted images and text for training has resulted in lawsuits (notably in 2023–2024). Legal outcomes will shape what training regimes are permissible.
Patent and inventor debates: Similar questions arise in patents when AI-generated inventions are claimed.

Ethical concerns

Attribution and credit: When AI contributes significantly to outcomes, how to recognize human vs machine roles? Transparency and provenance tracking are recommended.
Labor displacement: Creative professions may be affected; roles may shift toward curators, editors, and AI-guides rather than manual production.
Cultural appropriation and bias: AI models trained on biased datasets can reproduce problematic stereotypes or appropriate marginalized cultural artifacts without context.
Authenticity and deception: Deepfakes and synthesized creative artifacts can mislead audiences. There's a trade-off between generative power and trust.

Social effects

Democratization vs consolidation: Tools lower barriers to creative production (democratization) but major players (big tech) may centralize capabilities and economic gains.
Artistic ecosystems: AI may expand creative possibility spaces, spawn new genres, and alter the valuation of human-made work.

Governance and policy considerations

Standards for transparency: Labeling AI-generated content and providing provenance metadata.
Copyright reform: Clarify rights for AI-assisted works and obligations for dataset curation.
Support social safety nets and reskilling for workers in creative industries.

Practical governance proposals

Require model documentation—data provenance, training procedures, known limitations ("model cards").
Encourage human-in-the-loop designs where humans are credited for direction, curation, and final evaluation.
Develop sector-specific regulatory frameworks (e.g., advertising, journalism, education) to mitigate harms.

9. Current state-of-the-art and notable systems (as of mid-2024)

Language and text

GPT-4 family and other large multimodal models (capable of text + image) produce sophisticated narratives, poetry, and code; used in creative writing, ideation, and co-creation.
Specialized creative writing tools integrate LLMs for brainstorming, character development, and scene generation.

Image and multimodal

Diffusion models (Stable Diffusion, DALL·E 2/3, Midjourney) produce high-fidelity, stylistically diverse images from textual prompts.
Style mixing, inpainting, and latent traversals enable fine-grained creative control.

Music and audio

OpenAI Jukebox, Magenta’s models, and commercial tools can generate stylistically coherent music segments. Full long-form composition with human-level structure remains challenging.

Design and architecture

Generative design tools automate trade-off exploration for engineering constraints; AI suggests forms that humans refine.

Science and discovery

Tools like AlphaFold and generative chemistry models accelerate discovery. Debate continues on whether these systems are "creative" or powerful optimizers.

Human-AI co-creativity platforms

Many commercial products focus on co-creation: assistive interfaces where humans prompt iterations and select outputs (e.g., image generation pipelines, collaborative writing apps).

Empirical findings

Studies often show humans rate AI-generated artifacts as creative in constrained tasks; human editing increases perceived quality and creativity.
LLMs demonstrate associative flexibility and can perform well on some creativity tests (e.g., generating novel analogies), though they sometimes lack deep coherence over long compositions.

10. Future trajectories and research directions

Near-term (1–5 years)

Better controllability and interpretability: Tools to shape style, constraints, and intent; improved latent-space interfaces for designers.
Improved evaluation frameworks that combine automated metrics with standardized human evaluation protocols.
Legal clarifications and labeling norms for AI-generated art/content.
Wider adoption of co-creative workflows across industries.

Mid-term (5–15 years)

Systems capable of multi-step, conceptually coherent creative projects (e.g., multi-act plays, long symphonies with thematic development).
Greater blending of generative models with symbolic reasoning to produce artifacts with deeper causal or conceptual grounding.
Novel forms of creativity: AI may produce new aesthetics, genres, or hybrid media unforeseen today.

Long-term (15+ years)

If AGI-like systems or machines with richer models of human values and long-term goal-directed behavior arise, they might exhibit forms of transformational creativity currently unique to humans.
Profound societal shifts in how we value human vs machine creativity, new intellectual property regimes, and possibly reimagined economic structures for creative labor.

Open research questions

How to reliably evaluate transformative creativity?
Can models reason creatively across modalities and time (sustained conceptual development)?
How to ensure cultural diversity, fairness, and provenance in training data?
What architectures best support emergent creative capacities (hybrid symbolic-neural, multi-agent systems)?

11. Practical recommendations

For researchers

Explicitly define creativity measures and justify evaluation protocols.
Prioritize transparency: document datasets, training processes, and known biases.
Explore hybrid architectures combining deep learning and symbolic reasoning for richer conceptual creativity.

For creators and practitioners

Treat AI as collaborator: use it to generate ideation breadth, then apply human judgment for depth and context.
Maintain provenance records and be transparent about AI involvement in outputs.
Use AI to augment, not replace, domain expertise—especially where domain-specific constraints or ethics matter (e.g., medicine, journalism).

For policymakers

Encourage labeling and provenance norms.
Update IP frameworks to accommodate human-AI collaboration without disadvantaging creators or stifling innovation.
Support research into the social impacts of creative AI and fund reskilling programs.

For the public

Cultivate critical media literacy: be aware of synthetic content and its implications for trust.
Engage with new creative forms as opportunities for novel cultural expressions, while demanding transparency from platforms.

12. Conclusion

Can artificial intelligence be creative? The short answer: it already is, in important, meaningful senses. AI systems generate artifacts that are novel, surprising, and valuable by many standards and are increasingly integrated into creative processes across domains. However, the deeper philosophical question—whether AI possesses the subjective intentions, understanding, or agency associated with human creativity—remains unsettled and may not be necessary for many practical and artistic purposes.

A pragmatic view recognizes a continuum:

AI as a tool that extends human creativity (co-creative).
AI as a producer of outputs that communities can choose to value as creative (social attribution).
AI as an emerging agent whose creative capacities may grow with improvements in reasoning, long-term planning, and multimodal understanding.

Whichever conceptual stance one adopts, the important tasks are clear: develop robust evaluation methods, craft legal frameworks that fairly allocate credit and liability, mitigate harms (bias, misuse, labor impacts), and design AI systems that amplify constructive creative capacities while respecting cultural and ethical norms.