We stand at a peculiar crossroads in human creativity. An artificial intelligence can now compose a symphony indistinguishable from Beethoven, generate paintings that rival the masters, and write prose that moves readers to tears. Yet something in us recoils. The question haunting galleries, concert halls, and publishing houses is not whether machines can create—they manifestly can—but whether what they create deserves to be called art, and more pressingly, whether artists owe these systems anything at all.

The ethical fog thickens when we consider how these systems were trained. Millions of human artworks, poems, and compositions were fed into neural networks without explicit consent from their creators. A painter's distinctive brushwork, honed over decades, became data points. A novelist's unique narrative voice became statistical patterns. The machines learned to mimic not through study and struggle, but through wholesale consumption of human creative labour. This feels, to many, like theft dressed in scientific language.

Yet the counter-argument possesses its own force. Artists have always learned by studying predecessors. A young composer studies Bach; a painter copies the Old Masters in studios. We celebrate this as tradition, as the necessary foundation upon which new creativity builds. Is an AI's learning fundamentally different, or merely more efficient? Does efficiency transform inspiration into appropriation?

There is also the question of labour displacement. If AI can generate adequate design work, adequate marketing copy, adequate illustrations in seconds, what becomes of the thousands of professionals whose livelihoods depend on these skills? The answer cannot be merely technological optimism—the historical pattern of automation suggests that while new opportunities eventually emerge, the transition devastates communities unprepared for change.

Perhaps most troubling is how AI democratises creation while simultaneously devaluing the creator. Anyone can now produce visually striking images or grammatically perfect text. This abundance breeds indifference. When art requires no struggle, no sacrifice, no genuine human vision—when it can be summoned by typing words into a box—does it retain its power to move us? Art has always derived part of its significance from the knowledge that a human being laboured to create it, that they chose specific words or colours or notes because those choices mattered to them.

The ethical path forward likely lies not in prohibition but in transparency and consent. AI systems should be trained on works whose creators have agreed to this use, or compensated appropriately. Outputs should be clearly marked as machine-generated. We might even develop new aesthetic categories—not dismissing AI creation as illegitimate, but recognising it as something genuinely different from human art, with its own distinct value and limitations. The goal should be coexistence, not conquest.