Can editing be automated?

Hugh Farmar

Hugh Farmar

The existence of copy-editing programs such as Grammarly, which claims to find 10 times as many mistakes as word processors, and high-powered proofing tools such as Tansa, would seem to indicate it can.

Further, those unaware of the details of the investment research editing process would be forgiven for thinking it consists simply of a spell-check, liaison with compliance and hitting the Send button.

But while most good editors will have the required find-and-correct abilities possessed by the machine, they don’t simply open a document up and scan through it. They act first and foremost as one of an author’s most important readers.

This is because what a machine is unable to do is grasp the sense of a piece of work, respond in a human way to its tone and style, and formulate an opinion of how other people might receive it.

These skills suggest that editing is more art than mechanised procedure. It is about articulating concepts rather than merely checking facts and spelling.

No machine can spot when a point is laboured or when it is not given enough basis. No program can feel uncomfortable with a shift in argument and fix it by introducing a segue. Nor can they identify when a joke or reference might fall flat with the intended audience and suggest a punchy alternative. Not yet, anyway.

Editing can give research shape, polish and – most importantly for the crowded financial research marketplace – impact.