Summary
- A passage can affect us and still not deserve to be called authored in the same way.
- Not because using a tool is shameful, but because readers have a right to the same knowledge that so powerfully changed my participants’ judgements.
- The central question of this literary moment is not whether machines can produce sentences that move us.
That small shift deserves serious attention, because it is now an everyday experience in Pakistan. Students write essays with a chatbot open in another window. Young writers compete for attention in feeds saturated with machine-generated text. Editors, teachers and examiners must judge work when they can no longer be sure who — or what — authored it.
We are debating this issue in the wrong terms, swinging between panic on one side and dismissal on the other. Neither helps the reader in the middle, who simply wants to know what she is reading.
I came to this question through my own research. For my MPhil thesis in English Literature at Riphah International University, I asked 69 adult readers to examine six passages: three by human authors — Edgar Allan Poe, Rabindranath Tagore and Daniyal Mueenuddin — and three generated by machines working in dialogue with those traditions.
The participants first rated the passages without knowing their origins. Then they guessed who, or what, had written each one. Finally, the true authors were revealed, and they rated the texts again.
The first result should humble anyone who claims they can always spot the machine. Out of 138 authorship judgements, only 39.9% were correct. In 44.9% of cases, respondents honestly said they could not tell. At the level of the sentence, machine and human prose were, for most of my readers, indistinguishable.
The second finding matters more. Once the sources were revealed, the same words were judged very differently. Human passages were credited with far more conscious intention than AI passages — an average of 6.22 against 2.90 — and far more genuineness, 6.14 against 3.26.
Nothing in the prose itself had changed. What had changed was the story around the text. My participants were not only judging prose. They were judging a relationship: whether they believed a mind like theirs had suffered, chosen and meant something on the other side of the page.
It would be easy, and wrong, to read this as a simple victory for human writing. AI-generated passages did move readers before disclosure, and genuine emotion was reported for machine texts too. What collapsed after disclosure was not enjoyment. It was trust.
My readers drew a distinction our public debate usually misses, between being moved by a text and respecting its source. A passage can affect us and still not deserve to be called authored in the same way.
That distinction points to a practical rule, not just a theory. Authorship is a relation between the person who directs the work, the tools that generate it, the honesty with which its origins are declared, and the community that reads it.
One part of that relation cannot be handed to a machine: accountability. A machine cannot answer for a plagiarised paragraph, a defamatory claim or a fabricated fact. A human being must.
Where AI is used in writing — in universities, newsrooms, publishing houses and scholarship — disclosure should be the norm. Not because using a tool is shameful, but because readers have a right to the same knowledge that so powerfully changed my participants’ judgements.
In practice, disclosure need not be complicated: a brief note stating which model was used and for what, a methods line in a thesis, or an editor plainly asking a contributor what part of the work is theirs. The aim is not to punish tool use. It is to leave readers able to judge the work with full knowledge of its origins.
In Pakistan, this is not an abstract debate. If our only response is bans and detection software, we will teach concealment, not integrity. Students will learn to hide the tool instead of using it honestly, and examiners will remain uncertain about what they are marking.
The better policy for universities, editors and examiners alike is transparency: declare the tool, own the choices, and defend the words as yours.
The central question of this literary moment is not whether machines can produce sentences that move us. They plainly can. The question is whether we will stay honest about who wrote what — and whether we still believe that honesty matters.
My readers, once they knew the truth, clearly believed it did.
