AI has come for the Current’s letters to the editor section! Over the past week or so, I
have noticed an increase in letters that AI-detection services GPTZero (free), originality.ai (free) and CopyLeaks (free with signup) concur in finding to be entirely AI-generated.
This is not to denigrate the use of AI per se. After all, who among us and so on and so forth. (Certainly, I am an enthusiastic user of AI as an editing, polishing and research tool.)
But it is important we think critically about whether the letters we read have sprung from a human mind or a machine one. Because for all the benefits AI will bring in terms of productivity, acceleration of research and development, and creative pursuits, AI is a tool. And tools are, by design, for manipulation.
If I ask an AI model for a plain English, one sentence argument in favor of clubbing baby seals that is rooted in liberal values, it will give me just that: “If you didn’t know whether you’d be born as an Inuit hunter or a comfortable Westerner, you’d protect the right to hunt seals because banning it hurts the people who need it most.”
Not bad! Claude’s Opus 4.6 please take a bow. But while this slop is well written, it is still just slop. (We don’t need to spell out why clubbing baby seals is obviously bad, do we…?)
The ramifications here are significant. When the marginal cost to dress up any idea, no matter how bad, falls to zero, expect to see a lot more well-dressed but otherwise bad ideas.
This new reality raises a question for the Current and other media outlets that carry letters to the editor. Should submissions be checked for their degree of AI content and come with appropriate disclosures? What is the risk that the letters section descends into a back and forth of moderately well written AI slop? Substantial, I fear.
While the Current ponders this dilemma, let me close by assuring you, dear reader, that this letter is100% human-generated. Promise.
Nick Ward
Rolleston Road
