Are AI Art and Writing Truly Original?

If AI can write poems, paint portraits, and compose music… is any of it actually original?

It seemed like a straightforward question. It wasn’t.

What I found was a debate that’s currently raging in courtrooms, artist studios, and philosophy departments with no clear answer in sight. An AI-generated portrait wins a state fair art competition, and the art world erupts. An AI writes prose that moves people to tears, and writers ask: Is this creativity or imitation? A melody generated by an algorithm gets stuck in your head, and musicians wonder: where’s the line between inspiration and theft?

The deeper I dug, the more I realized this isn’t just about AI. It’s about what we mean by original in the first place, and whether human creativity is as mysterious as we would like to believe.

What Does Original Actually Mean?

Before we dive into AI, let’s get honest about human creativity for a moment. When Picasso said, “Good artists copy, great artists steal,” he wasn’t confessing to plagiarism. He was revealing an uncomfortable truth about how all creativity works.

Every artist, writer, and musician learns by absorbing existing work. You read thousands of books before writing your first novel. You study masterpieces before picking up a paintbrush. Your influences seep into your work whether you realize it or not. Bob Dylan borrowed folk melodies. Shakespeare “borrowed” plots from earlier writers. Led Zeppelin… well, let’s just say they were very “inspired” by old blues records.

So what makes something original? Most definitions revolve around a few key ideas:

1. Novelty – It’s new or unique in some meaningful way
2. Independent creation – It wasn’t directly copied from something else
3. Creative input – It reflects choices, interpretation, or expression from the creator

Notice what’s not required: creating something from absolute nothingness. That standard would disqualify virtually all human creativity throughout history.

How AI Actually Creates

Here’s where things get fascinating. Modern AI systems like GPT-4, Midjourney, or DALL-E don’t store a database of images or text to copy from. Instead, they learn patterns.

Take it like this: When you learn a language, you don’t memorize every possible sentence. You internalize grammar rules, vocabulary, and patterns of expression. Then you generate new sentences you have never heard before. AI does something conceptually similar, just with billions of parameters trained on vast datasets.

For text models: They learn statistical relationships between words and concepts. When you prompt “write a story about a lonely robot,” the AI doesn’t retrieve a pre-written robot story. It generates text token by token, predicting what comes next based on patterns it learned, filtered through your specific prompt.

For image models: They learn the relationship between text descriptions and visual features. When creating an image of “a Victorian mansion at sunset,” they are not photoshopping existing mansion pictures together. They are generating pixels based on learned associations between concepts and visual patterns.

The key question: Does learning from patterns make the output non-original? Because if so, we have a problem. That’s exactly how human creativity works, too.

Also Read: Newsrooms Have a New Colleague: AI. Now What?

The Divine Spark Argument

Some argue that AI lacks the mysterious human element—consciousness, intentionality, lived experience—that makes true originality possible. Something is appealing about this idea. When a human artist paints their grief, writes about lost love, or composes music reflecting their joy, they are channeling genuine experience and emotion.

AI has no grief to channel. It doesn’t understand sunset’s beauty or loss’s ache. It manipulates patterns without comprehension.

But here’s the counterpoint that might make you uncomfortable: Does the audience care?

When Jason Allen’s AI-generated artwork “Théâtre D’opéra Spatial” won the Colorado State Fair in 2022, people responded emotionally to the image before knowing its origin. The artwork evoked feelings, sparked imagination, and demonstrated aesthetic merit. Did it become less meaningful once its AI origin was revealed?

And consider this thought experiment: If I write a moving poem about war despite never having fought in one, is it less original than a veteran’s poem? If a deaf composer like Beethoven creates a symphony he cannot hear, does the lack of auditory experience make it less creative?

The “divine spark” argument struggles because we can’t actually prove humans have something AI categorically lacks. We assume we do because… well, we are human, and it feels true. But “it feels true” isn’t a rigorous philosophical argument.

The Legal Perspective: Copyright and Originality

The law offers surprisingly clear guidance, even if you might not agree with it. In the United States, copyright requires a work to be an “original work of authorship.” Courts have consistently held that this requires:

1. Independent creation (not copied)
2. Minimal creativity

The bar for “minimal creativity” is remarkably low. Arranging names in a phone book alphabetically? Not creative enough. Arranging them in a creative order? Original enough for copyright.

Here’s where AI gets legally complicated:

The U.S. Copyright Office has stated that AI-generated works without human creative input cannot be copyrighted. In 2023, they denied copyright for images created by Midjourney because “the Office will not register works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from a human author.”

But notice the caveat: “without any creative input from a human author.”

If you spend hours crafting the perfect prompt, curating outputs, editing and selecting from hundreds of generations, and integrating AI elements into a larger human-created work, courts may well find human creative input sufficient for originality. The legal landscape is still evolving.

The Training Data Dilemma

Here’s where things get ethically thorny. AI systems trained on billions of images and texts often didn’t ask permission from the original creators. Artists discovered their work in training datasets without consent. Writers found their books used to teach AI without compensation.

The argument against originality: If AI learned from copyrighted works without permission, can its outputs be original? Isn’t it just recombining stolen materials?

The counter-argument: Human artists also learn by studying copyrighted works without paying licensing fees. We don’t require painters to pay every artist whose museum work they studied. The question is whether AI learning constitutes copyright infringement or fair use—a question currently being litigated in multiple lawsuits.

Several ongoing cases will shape this debate:

• Artists are suing Stability AI, Midjourney, and DeviantArt
• Authors including Sarah Silverman are suing OpenAI
• Getty Images is suing Stability AI over image training data

The outcomes will significantly impact how we understand AI originality from a legal and ethical standpoint.

What Makes AI Different (If Anything)

Let’s be intellectually honest about what distinguishes AI from human creativity:

Scale and speed: AI can generate thousands of variations in minutes. This quantity doesn’t necessarily equal quality, but it changes the creative landscape significantly.

Lack of intentionality: When AI creates, it has no goal beyond pattern completion. It doesn’t intend to convey meaning, evoke emotion, or communicate ideas. Any meaning exists only in the human creator’s prompt and the audience’s interpretation.

No lived experience: AI hasn’t experienced life, felt emotions, or navigated the world. Its “understanding” is purely linguistic and visual pattern recognition.

Transparency of process: We can theoretically examine AI’s training and generation process. Human creativity remains largely mysterious, even to neuroscientists.

But here’s what’s similar:

Both learn from existing works. Both combine learned patterns in new ways. Both can produce novel outputs not seen in training. Both can surprise their “creators” (the AI’s user or the artist’s conscious mind)

A More Useful Question

Instead of asking “Is AI art truly original?” maybe we should ask: “What kind of originality does AI offer, and how does it compare to human originality?”

This reframing acknowledges that originality isn’t binary. It exists on a spectrum with different flavors:

Combinatorial originality: Taking existing elements and combining them in new ways (arguably what both humans and AI do)

Expressive originality: Conveying a unique personal perspective or experience (currently human territory)

Technical originality: Novel methods or approaches (AI excels here in certain domains)

Conceptual originality: New ideas or ways of seeing (requires prompting from humans, but AI can help explore conceptual spaces)

The Uncomfortable Truth

Here’s what might be the most challenging takeaway: AI has revealed that much of what we called “creativity” is pattern recognition and recombination, skills that can be replicated algorithmically.

The work that AI handles well—generating competent prose, creating appealing images in established styles, composing pleasant melodies—perhaps wasn’t as mysteriously “creative” as we thought. These skills, while valuable, may be more mechanical than we wanted to admit.

What AI struggles with reveals what’s genuinely unique about human creativity:

• Intentional meaning-making rooted in experience
• Understanding context and subtext deeply
• Creating work that connects to shared human experience
• Making creative choices that reflect values and worldview
• Innovating conceptually, not just aesthetically

AI doesn’t replace human creativity; it forces us to understand what human creativity actually is.

So… Are They Original?

If originality means “created from nothing with zero influence from existing work,” then no, but neither is human art.

If originality means “independently generated based on learned patterns without direct copying,” then yes, AI outputs can be original in the same way human creations are.

If originality requires conscious intentionality and lived experience, then no AI lacks the human element that makes creativity meaningful.

The practical answer: AI art and writing exist in a new category that challenges our existing definitions. They are not original in the same way human work is, but they are not simply non-original either. They are computationally generated based on learned patterns, guided by human prompting, capable of novelty but lacking intentionality.

Instead of forcing them into existing boxes labeled “original” or “derivative,” perhaps we need new language for new kinds of creation.

What we can say with certainty: AI has changed the conversation about creativity forever. It’s exposed assumptions about originality we didn’t know we had, raised copyright questions we haven’t resolved, and forced us to articulate what makes human creativity special.

And ironically, in trying to replicate human creativity, AI has helped us understand it better than ever before.

What do you think? Does knowing something was AI-generated change how you experience it? Does originality require consciousness, or just novelty? The debate is far from settled, and maybe that’s exactly as it should be.