A Writer's Identity Crisis in the Age of AI: Arundhati Roy's Fears and the Copyright Fight

In just a few minutes, artificial intelligence can now produce a full feature, story or essay on demand. Ask for a piece that is slow, deep and close to life like an Abbas Kiarostami film, or one written in the literary manner of a familiar Nobel laureate, and AI can arrange the words to order almost instantly.

Technology has put creativity within everyone's reach, but that ease does not always bring good — especially for writers. If a writer's own language, experience, feeling and distinctiveness become overshadowed by machine imitation, literature can lose its original voice, and technology becomes not a helper of creativity but a silent tragedy for the writer's identity.

The year was 2016, and manuscripts were being screened for Japan's Shinichi Hoshi literary prize. A total of 1,450 manuscripts had been submitted, and 11 were shortlisted in the preliminary round. One of them was "The Day a Computer Writes a Novel."

The real surprise lay elsewhere. Just before the winner was to be announced, the judges learned that the novel was not written by a human at all — it was written by an AI program called "Sekai Camera." The incident caused a considerable stir in the literary world, and in the end the AI-written novel was removed from the competition.

Since then, the use of AI in writing has been fiercely debated. There have even been reports that an AI-written work was nominated for the Nobel Prize in Literature — although how true that is remains unconfirmed.

In the age of AI, a writer's greatest crisis is not merely competing with technology but holding on to their own distinctiveness. A writer builds a personal style over years of experience, observation and practice. But when AI can imitate that language, word choice and manner of narration, the very definition of originality comes into question — and if a writer's own writing is suspected of being AI imitation, the burden falls on them to prove their creation is genuine.

Celebrated author Arundhati Roy has spoken openly about this crisis. On AI, she says she fears that very soon, when she writes in her own way — in the language she has spent decades making entirely her own — she will be told: this does not sound like your writing; this is actually AI.

At the centre of Roy's anxiety lies the question of a writer's creative labour and intellectual property. Over years of experience, thought, emotion and observation, a writer forms a distinctive style; their word choice, sentence construction, manner of description and tone become part of their identity.

But when that very writing is used, without the writer's permission, to train large language models, the question arises: who owns a writer's creative assets? And once the technical ability exists to imitate their language and style, who bears the ethical and legal responsibility for that imitation?

Roy described her own experience while raising the writers' copyright lawsuit against Anthropic. According to her, she is one of the authors whose books were used to train large language models without permission, and she too was sent a notice following a collective lawsuit over the matter.

In Roy's account, the notice proposed settling the dispute by paying roughly two thousand dollars per book. Her complaint is that the long years, labour and life experience spent creating those books cannot be reduced to a sum of money in this way. She also noted that the notice said the books had been downloaded from two websites that host pirated PDFs — while also stating that Anthropic admitted no wrongdoing.

The metaphor Roy used to describe the whole experience is especially striking: it felt, she said, as if limbs were being cut from her body without her knowledge — because the machine trained on her writing had already used her creative labour in the process of its making.

The crisis is not limited to world-famous authors. Sarah Suzuki Harvard, a 32-year-old copywriter in Brooklyn, New York, is not accustomed to using very emotional language professionally. But amid AI's growing influence, she is changing her familiar writing style to make her work recognisable as human — deliberately using colloquial language in posts and essays, sometimes adding conversational expressions.

All told, it is hard to label AI's use in writing simply positive or negative. For research, gathering information, editing or finding the seed of a new idea, AI can be an effective aid. But using a writer's work for training without consent, and producing new writing in imitation of that creativity, raises ethical questions. Banning the technology may not be the answer; what is needed is transparency, copyright protection and guaranteed consent — because literature is not merely an arrangement of words, but is bound up with human experience, feeling, responsibility and philosophy of life.