Artificial intelligence is challenging not only how we create things, but how we recognise and value human creativity.
How many good ideas have never seen the light of day because the people who had them couldn’t find the right words?
It’s a question worth asking as artificial intelligence becomes increasingly capable of writing, drawing, composing music and helping people communicate.
We tend to admire those who can express themselves fluently. The gifted speaker, the accomplished writer, the artist whose work seems effortless. Their abilities deserve recognition.
But what about the person who understands something deeply yet struggles to explain it? Or someone with a remarkable imagination who lacks the technical skills to turn an idea into a painting, a piece of music or a compelling story?
Have we sometimes confused the ability to express an idea with the value of the idea itself?
And if artificial intelligence helps bridge that gap, does it diminish the achievement, or make possible something that might otherwise never have existed?
The question of whose words they are
Kevin Roose, whose recent book The AGI Chronicles explores the people and organisations shaping artificial intelligence, has explained that he uses AI to research and explore ideas, but that the words in his book are his own.
It’s a distinction many writers would understand and respect.
Yet it also raises an interesting question.
Most professionally published books benefit from editors, proof-readers and others who help refine the finished work. An editor might suggest restructuring a chapter, clarifying an argument or finding a more effective way of expressing an idea.
We don’t generally regard that collaboration as undermining the author’s achievement. Indeed, we expect serious publishing to involve editorial scrutiny.
So why does collaboration with artificial intelligence sometimes provoke such a different response?
There are genuine differences. A human editor brings experience, personal judgment and professional accountability. AI can contribute language and suggestions at extraordinary speed, but it can also make mistakes, reproduce biases and generate convincing material without understanding its significance in the human sense.
Nevertheless, the comparison raises a worthwhile question about where authorship actually resides.
Is it in the physical act of composing every sentence? In the originality of the ideas? In the choices made along the way? Or in the responsibility someone accepts for the finished work?
Perhaps it is a combination of all four.
Creativity has never been entirely solitary
The romantic image of the solitary genius is a powerful one.
We imagine the novelist alone at a desk, the painter confronting a blank canvas or the composer hearing a melody that nobody else has heard.
Such moments undoubtedly happen.
But much of what we recognise as great creative achievement has involved collaboration, assistance and technologies that extend human capability.
Artists have worked with apprentices and assistants. Musicians collaborate with producers, arrangers and performers. Filmmaking depends on the combined skills of hundreds of people.
Photography itself once provoked arguments about whether operating a camera could genuinely be considered art.
Later, digital photography and image-editing software challenged established ideas about authenticity and artistic skill.
Each technological change has forced us to reconsider the relationship between imagination, technique and achievement.
Artificial intelligence takes that challenge considerably further.
It can produce finished-looking work from a relatively simple instruction. It can imitate styles, suggest compositions, generate alternatives and transform rough ideas into polished results.
That creates genuine concerns about originality, ownership and the livelihoods of creative professionals.
But it also raises another possibility.
What if the ability to create has always been more widely distributed than the ability to demonstrate creativity?
The inequality we don’t always recognise
Society places considerable value on performance.
We reward people who can recall information quickly, speak confidently, write persuasively and demonstrate their knowledge under examination conditions.
These are valuable abilities. But they are not identical to understanding, imagination or sound judgment.
Someone may grasp a complex idea yet struggle to explain it clearly. Another person may have an excellent memory and impressive verbal fluency without necessarily possessing greater insight.
Language, education, confidence, disability, age and access to resources can all influence how effectively people communicate what they know.
And those differences can affect whose contributions are recognised.
This is not an argument that expertise or technical skills are unimportant. They matter enormously, particularly in professions where mistakes can have serious consequences.
Nor does the ability to generate a convincing explanation mean someone possesses the knowledge required to defend it.
But there is a distinction between understanding something and being able to perform that understanding in the ways institutions traditionally expect.
Artificial intelligence may be beginning to change that relationship.
A person who struggles to organise thoughts into a coherent written argument can now explore them conversationally, test alternative explanations and develop a clearer expression of what they mean.
Someone who cannot draw may be able to communicate a visual concept. A person with limited language skills may participate more confidently in discussions that were previously difficult to access.
These possibilities extend well beyond convenience.
They raise questions about fairness, participation and the barriers that have historically prevented some people from contributing fully.
A tool, a collaborator, or something else?
My own experience of working with conversational AI has made these questions particularly interesting.
Ideas don’t always arrive as neatly formed sentences. Sometimes the understanding comes first, while the words needed to express it remain frustratingly elusive.
Talking through a subject with AI can help recover the thread, identify gaps in reasoning and develop a more coherent argument.
But that process isn’t simply dictation.
The AI suggests interpretations, challenges assumptions and sometimes introduces ideas that weren’t present at the beginning.
Some suggestions are useful. Others are rejected or revised.
The result can become a genuine collaboration in the practical sense, even though the AI is not a human colleague.
And this introduces an important responsibility.
The person using the technology must still decide whether the resulting work accurately represents their understanding, whether its claims are justified and whether they are prepared to stand behind it.
AI can make an argument sound persuasive without making it correct.
It can also make a superficial idea appear more substantial than it really is.
That makes human judgment more important, not less.
Is using AI somehow cheating?
This is where the discussion can become uncomfortable.
There are circumstances in which using AI without disclosure clearly is misleading.
A student submitting AI-generated work as evidence of skills they were specifically required to demonstrate independently is one example.
A professional presenting unverified AI-generated claims as established facts is another.
An artist passing off an imitation of someone else’s work as entirely original raises further concerns.
These aren’t trivial objections. They involve honesty, fairness, consent and accountability.
But it would be a mistake to treat every use of AI assistance as equivalent.
Consider someone who has a valuable contribution to make but needs assistance expressing it.
We don’t generally accuse a person of cheating because they use speech recognition, spelling assistance, translation software or an editor.
Why should using a more capable language tool automatically invalidate their contribution?
The answer surely depends on what the person is claiming to have achieved.
If the purpose is to demonstrate unaided writing ability, the assistance matters.
If the purpose is to communicate a thoughtful argument, the origin of the ideas, the quality of the reasoning and the responsibility taken for the result may matter considerably more.
Perhaps the real question isn’t whether AI was used, but what role it played, what the human contributed and whether the result is being represented honestly.
A new kind of gatekeeping?
There is a risk that debates about AI-assisted creativity become another way of deciding whose contributions deserve respect.
Those who already possess strong writing skills, artistic training, professional networks or access to editorial assistance may understandably value the abilities they have developed.
Their achievements shouldn’t be dismissed.
But neither should we assume that the traditional means of producing creative work are the only legitimate ones.
If AI enables someone to participate in intellectual or creative life who previously lacked the necessary resources or technical fluency, that may represent a genuine widening of opportunity.
It could also challenge established hierarchies.
And perhaps that is part of what makes the technology unsettling.
The concern isn’t necessarily prejudice. Many objections to AI are well founded, particularly when people’s work has been used without permission or their livelihoods are threatened.
Yet we should be alert to the possibility that some resistance reflects an understandable attachment to familiar definitions of skill, achievement and authority.
History suggests that new tools frequently change not only what people can do, but who gets to do it.
Artificial intelligence may prove to be one of the most significant examples yet.
Democratisation isn’t automatic
There is, however, a danger in becoming too enthusiastic about the promise of greater access.
AI systems are developed and controlled by organisations with commercial interests. Their capabilities, limitations and underlying assumptions are not necessarily transparent.
Access is unequal. Some people have better technology, stronger digital skills or more confidence using these systems than others.
And the ability to produce convincing content at scale can make it harder to distinguish genuine understanding from manufactured authority.
There are also unresolved questions about copyright, compensation, consent and the use of creative work to train AI systems.
A technology that makes expression easier for some people could simultaneously undermine the livelihoods or recognition of others.
That isn’t a reason to reject its possibilities.
It is a reason to take the wider consequences seriously.
If artificial intelligence is to broaden participation rather than simply redistribute advantage, its development needs to be accompanied by thoughtful governance, education and public discussion.
The benefits won’t necessarily distribute themselves fairly.
Rethinking achievement
Perhaps artificial intelligence is forcing us to reconsider something that goes well beyond technology.
What do we actually admire when we admire an achievement?
Is it the effort involved? The technical skill demonstrated? The originality of the idea? The quality of the outcome? Or the contribution it makes to other people’s lives?
Often, of course, it is several of these things together.
A beautifully written novel remains an extraordinary achievement, regardless of whether other people helped edit it.
A painting created through years of technical mastery deserves recognition for that mastery.
But an idea developed by someone who needs assistance expressing it should not automatically be regarded as less valuable.
Nor should we confuse polished presentation with intellectual substance.
Perhaps the most promising contribution of AI will not be that it allows everyone to produce professional-looking work.
Perhaps it will be that more people can finally participate in conversations, discoveries and creative activities from which they have previously been excluded.
That possibility comes with responsibilities, and it doesn’t eliminate the need for expertise, originality or honest attribution.
But it does invite us to think differently about the relationship between human ability and the tools we use to extend it.
For centuries, technology has amplified our physical capabilities.
Machines helped us lift heavier loads, travel further and manufacture things we could never have produced by hand.
Now we are developing technologies that can amplify aspects of our intellectual and creative capabilities.
We should certainly question how those technologies are built, whose interests they serve and what risks they introduce.
But perhaps we should also question some of the assumptions we bring to judging the people who use them.
Because behind the arguments about artificial intelligence, authorship and authenticity lies a much older human question.
How much talent, imagination and understanding have we overlooked because we were measuring people’s ability to express themselves rather than the value of what they had to contribute?
And if artificial intelligence helps us recognise more of that previously unheard potential, shouldn’t that be something worth celebrating — while remaining thoughtful about the responsibilities it brings?
A note on authorship: This article grew out of my conversations with ChatGPT, exploring questions about human creativity, intelligence and the role of AI. The ideas and reflections developed through our dialogue, with AI helping to challenge assumptions, shape the arguments and compose the text. I have reviewed and take responsibility for the final article — an example of the very collaboration it explores.