AI

Letter to the editor(s): Breaking the curse of the AI editor

Article by:Alex Maxwell

Share:

A 10% chance “AI could kill all humans” in the next decade. These are the words of a senior researcher at Anthropic that have led to this week’s news being dominated by doomongering prophecies of humanity’s extinction at the digital hands of AI. 

Yet, in spite of these warnings, another story released on the same day also showed how a growing number of companies are actively favouring employees for promotions based on their use of AI, regardless of their performance (which, to me, seems totally nonsensical and more of an attempt to attain/force an ROI).

Aside from the obvious extinction fears, the problem with these prophecies and workplace pressure is that they portray an image of AI as an omniscient source – a deity, almost – that means people treat its ‘opinion’ as gospel. But while AI is starting to excel at computational, mathematical and analytical endeavours, where the doomsday fears derive from, its impact in creative areas like writing has been less effective. So much so that big tech companies have resorted to buying and then scandalously burning swathes of secondhand books to improve this skill. 

Despite these flaws, the depiction of AI and employer pressure is leading people to use it for everything, including editing content requiring human opinions and intuition. And it’s a problem we’re encountering more frequently. 

In one instance, an article we wrote for review was fed through an AI model to create an AI version of the article and the AI version of the article was then reviewed by another AI model before human feedback subsequently commented on the changes and phrasing choices that the AI had made to the original piece forming a cycle of AI on AI and human on AI that made the process far more convoluted and protracted than if a human opinion had simply been applied to the original piece. 

You get the picture. 

Even if they had used AI to review the original piece, that would have been better. But still, that’s overriding the fact the article is meant to be reflecting a human’s personal experience and expertise. This seems a harder and harder point to get across, to instil in people’s minds. 

It’s worth saying that this blog could also easily be titled the curse of too many editors. These are pitfalls that exist in human processes too. But relying solely on an AI editor, like having too many editors, can strip out the personality, context and phrasing that might’ve landed with the person who is authoring the piece – by the time they see it, this substance has been lost, and a lengthy rework ensues.

But the real message in all this is far more straightforward: if you’re editing an opinion piece, apply your own opinion to it first. If you then feel a need to gauge AI’s verdict, fine, but it should be supplementary to your own thoughts. 

In the face of relentless AI pressure, what leaders and marketing teams need to do is break free from the AI editing curse, of trusting its position blindly. What people want to hear is what you have to think.

Share:

Back to blog