Meru Gokhale@MeruGokhale
How tempting it is to join a pile-on.
Especially when the target is @GrantaMag, and the mistake looks obvious after someone else has found it.
Granta is now caught up in the latest AI-writing controversy. A story published on its site is under review after claims that it was AI-generated. People have run it through detectors. They have pulled out the tics. They have laughed at the author's prose, the judges, the editors.
The accusation is: how could they not know?
And the subtext is: I would have known.
Maybe. Maybe not.
If an undisclosed AI entry won a human-only prize, investigate it properly. Writers who submitted original work deserve that. Prizes need rules they can enforce, and magazines need processes that do not collapse the moment suspicion appears.
Laughing at the failure is too easy.
Work written with or by AI is going to keep arriving on editors’ desks. In submissions, manuscripts, prize entries, blurbs, reports, author notes.
Editors are are paid to protect the author's work. Doing that properly now requires AI literacy.
If you choose to use AI, you need to know how to use it with discipline: what to ask, what to check, when it is flattening a voice, when it is inventing.
If you do not use AI in your own edits, you still need to understand it. Learn the syntax. Learn the habits. Learn what detectors can and cannot tell you. Learn how to investigate suspicion without humiliating a writer first.
If an editor chooses not to learn this, they are putting their authors at risk.
Most editors I meet want to learn. I have run workshops with experienced editors who ask careful, practical questions because they know the work has changed.
If we turn every mistake into a witch hunt and public shaming, people will stop admitting what they do not know.
And for a crowd that talks so much about what it means to be human, there is a simpler test available here: behave humanely towards the author and editors at the centre of this.