Open-Ended AI: Can Machines Ever Be Truly Creative — or Just Useful?

By Stuart Kerr, Technology Correspondent — LiveAIWire

Published: December 1, 2025 | Updated: December 1, 2025 • Contact: liveaiwire@gmail.com

When an AI paints a surreal landscape or composes a haunting melody in seconds, it feels like witnessing magic. But is it true creativity — or just clever imitation? As generative AI tools flood the creative landscape, this question looms large. Can a system with no emotions, no experiences and no consciousness truly “create”? Or is AI destined to remain a sophisticated assistant — powerful, but never truly original?

What we mean by “creativity”

Scholars commonly define creativity by three core criteria: novelty, surprisingness, and value (or usefulness). ResearchGate+2compass.onlinelibrary.wiley.com+2

• Novelty — producing something new.

• Surprisingness — deviating from expectation in a meaningful or insightful way.

• Value — being judged worthwhile by a community (aesthetically, emotionally, functionally).

These criteria form a baseline against which both human and artificial creators tend to be measured. ResearchGate+1

With generative AI — whether image, music or text — we increasingly see outputs that satisfy these criteria. From digital artworks to short stories, the question is no longer whether AI can produce “something different”, but whether that “something different” qualifies as creativity. OUP Academic+2ResearchGate+2

Evidence that AI can approximate creativity

AI-generated works already match many creative benchmarks

Recent studies document that AI-generated outputs are capable of producing artefacts historically considered “creative”: paintings, music compositions, stories. OUP Academic+1

For instance, a 2024 empirical study showed that generative-AI assistance caused stories to be rated by human judges as more creative, better written and more enjoyable — especially when authored by people who might otherwise struggle with narrative writing. Science

A systematic meta-analysis published in 2025 surveyed 28 studies with over 8,000 participants: while generative AI alone did not outperform humans statistically in creative tasks, collaboration between humans and AI significantly boosted creative performance (effect size g ≈ 0.27). arXiv

AI as a creativity amplifier rather than replacement

AI seems particularly effective as a creative multiplier — a tool that amplifies human creativity rather than replaces it. That’s the conclusion of a growing body of research suggesting human-AI collaboration yields stronger results than either working alone. arXiv+2ResearchGate+2

In practical terms, artists, writers and creators increasingly view AI as a “muse” or co-pilot — able to suggest ideas, generate rough drafts, remix influences — while humans retain final judgment, emotional nuance and intent. Cornell Tech+2blog.jlipps.com+2

But there are deep challenges — and arguments against “true” machine creativity

Lacking consciousness, intention and emotional authenticity

Critics argue that no matter how sophisticated, AI lacks consciousness, lived experience, emotional reality and intentionality — dimensions many consider essential to “real” creativity. Nature+2SpringerLink+2

A 2025 philosophical critique demonstrates that while AI outputs can satisfy novelty/usefulness, the absence of cognition and intentionality places a boundary: creativity without subjective experience may be possible functionally, but not existentially. SpringerLink+1

Risk of homogenisation, flattening and over-reliance

Generative AI is trained on huge datasets reflecting past human work. As a result, it tends to remix, rehash, or imitate — which risks converging on familiar patterns and tropes, even when those patterns are technically “novel.” indiaai.gov.in+2sciencedirect.com+2

In some evaluations, audiences still preferred human-made art over AI-made alternatives — especially in perceived emotional depth or perceived value, even when blind-tested. ResearchGate+2Harvard Gazette+2

Moreover, AI’s ease and speed can tempt creators to rely heavily on it, potentially eroding human creative skills or motivation: after all, why labour long-hours in creative struggle if a machine can draft a passable version in minutes? aokistudio.com+1

Ethical, cultural and value-based limitations

Beyond aesthetics, art and creativity have cultural, historical, emotional, moral dimensions. Machines — lacking lived context — may struggle to embed real meaning, cultural memory, ethical tension or unique perspectives. Some critics claim this leads to an artistic “flattening” or a form of homogenised, emotionally shallow mass output. Harvard Gazette+2SpringerLink+2

There’s also the question of authorship and ownership: who owns the creative output when a machine plays a central role? And does mass-produced AI art cheapen the value of human creativity and labor? indiaai.gov.in+2ojs.bonviewpress.com+2

A nuanced view: degrees of creativity & hybrid intelligence

The debate is rarely binary. Many researchers and theorists suggest thinking of creativity as a spectrum — where some tasks (craft-level creativity, remixing, ideation) are well within AI’s reach, while others (visionary art, emotionally resonant literature, context-rich cultural commentary) remain strongly human. sciencedirect.com+2SpringerLink+2

Under this framing:

• AI can excel at combinatorial or explorative creativity — mixing styles, recombining forms, rapidly iterating variants.

• Humans remain critical for transformational creativity — injecting subjective vision, emotional nuance, ethical depth, cultural meaning.

And there’s value in hybrid creation workflows, where tools and humans complement each other — machines handling scale, brute-force ideation, iteration; humans providing soul, context, judgement. Many creators already work this way. Cornell Tech+2OUP Academic+2

What this means for creators, culture and the future of art

• For creators and artists: Treat AI as a powerful tool — but don’t forget your role as curator, storyteller, value-maker. Use AI to prototype, remix, experiment — but bring human insight to shape meaning.

• For cultural institutions and society: Establish standards and criteria for attribution, transparency, and ethical use. As AI-generated works proliferate, communities will need frameworks to assess originality, value, rights.

• For audiences and consumers: Stay critical. Recognise the difference between novelty and depth, between surface-level aesthetics and real emotional or cultural resonance. Appreciate AI outputs — but don’t conflate quantity with creativity.

• For AI developers & researchers: The next frontier isn’t just better output — it’s richer processes. Building AI that better understands context, culture and long-form coherence may help bridge the gap from imitation to inspiration.

Conclusion: AI’s creative capacity is real — but limited

The fast-advancing world of generative AI proves one thing clearly: machines can produce work that looks creative, surprising, useful — often convincingly so. In many domains, AI already functions as a creative assistant, idea generator, draft maker.

But creativity is more than output: it’s insight, emotion, context, memory, struggle, risk. These remain — for now — largely human domains.

So yes, AI can be creative — in degrees, for certain tasks. But true, full-spectrum creativity — the kind that moves hearts, challenges minds, reshapes culture — remains, for now, our territory.

In the end, the question isn’t just whether machines can create — but whether we use them to expand creativity, or outsource it entirely.

© LiveAIWire 2025 — Supplemented by AI and Caffeine

About the Author

Stuart Kerr is a correspondent on AI at LiveAIWire. He reports on how AI reshapes work, media and the systems people rely on.