Everyone has copied everyone at some point. That is not some new phenomenon invented by AI. Artists copied the masters before them; students copied techniques from teachers; apprentices reproduced the work of established artists; people studied paintings by recreating them, tracing outlines, copying compositions and learning by imitation. We literally did versions of this in art lessons at school.
The same applies to writing and information. People copied passages from books, photocopied articles, transcribed information, collected newspaper clippings, quoted other people's work, borrowed ideas, adapted arguments and built new work on top of material produced by previous generations. Before photocopiers there were carbon copies. Before that there were scribes copying manuscripts by hand. Before that, knowledge was transmitted orally and through imitation.
The important point is that copying has always been part of how human culture develops. Almost nothing exists in an intellectual vacuum. Artists study other artists. Writers read other writers. Scientists build on previous research. Engineers examine existing designs. Musicians borrow structures, rhythms and techniques. Students learn by reproducing what somebody else has already demonstrated before they develop their own variation.
Technology has repeatedly made this process faster and easier.
The printing press massively accelerated the reproduction of information. Photography made visual reproduction vastly easier. The photocopier allowed people to duplicate documents almost instantly. Tape recorders, samplers, scanners, digital photography, file sharing and the internet progressively reduced the cost and friction of copying.
And now AI has taken another step in that progression.
That does not mean that every use of AI is automatically legitimate. There are genuine questions about copyright, attribution, licensing, consent, compensation and whether particular systems have been trained or used in ways that violate existing rights. Those are perfectly reasonable arguments to have.
But pretending that AI somehow introduced the fundamental concept of learning from, reproducing, transforming or being influenced by previous human work is historically absurd.
AI did not invent imitation.
It industrialised and accelerated it.
That is arguably the uncomfortable part of the debate. For centuries, humans could copy things, but the process was constrained by time, labour, skill and access. You had to sit in front of the painting. You had to read the book. You had to copy the passage. You had to trace the drawing. You had to photocopy the pages. You had to learn the technique.
AI dramatically reduces those constraints. And that is why the argument feels so different now. It is not necessarily because the underlying human behaviour is completely unprecedented; it is because the scale, speed and automation are unprecedented.
The same technological pattern has appeared repeatedly throughout history. A capability becomes cheaper and faster, the old economic model is disrupted, the people whose work depended upon scarcity become angry, and society then spends years arguing about where the boundaries should be.
So yes, have the copyright debate. Have the compensation debate. Have the attribution debate. Establish rules where genuine rights are being infringed.
But portraying AI as some uniquely evil invention because it learns from existing human-created material requires conveniently forgetting how human civilisation has actually worked for thousands of years.
We have always stood on the shoulders of previous generations.
The difference now is that the machine can climb the ladder considerably faster than we can.