As someone who writes for a living, I should loathe A.I. 

Large Language Models have made, and will continue to make, many traditional writing and editorial jobs defunct. A trained fact-checker needs a salary; Perplexity doesn’t. Gemini can render a press release in seconds. It might not be good, of course, but it will be done, and the writer responsible for the prompt will have been relegated to the status of sub-editor. Chat GPT has succeeded in turning the M-dash into a mark of suspicion. 

So why do I use A.I. every day?

Because A.I. is just like any other technology: disruption is soon followed by a period of adjustment. Once things settle down, society becomes unimaginable without the technology in question. So it was with the printing press, the camera, and the telephone. Luddites and cheerleaders are both equally wrong. The genie can’t be rebottled. But nor can the costs be ignored.

A.I. will certainly continue to pose moral, aesthetic, and economic challenges. Meanwhile it is important for writers to pitch themselves at the far end of its curve. 

I do this in my own work by drawing strict personal parameters. Everything I write is, as it were, organic. After that, I find LLMs immensely useful in rooting out logical and structural imperfections. Far from usurping my creative work, A.I. augments my ability to perform that work with greater scope and efficiency. It may get things wrong – sometimes basic factual things – but that only means that I am obliged to follow up its claims and suggestions with my own research. A.I. can pose an idea – a strapline or a headline, say – but only I can judge its worth because only I have a real relationship with the affected stakeholders. And learning how to prompt an LLM effectively is itself a good way of improving one’s writing and organising one’s thoughts.

A.I. is not to be feared. Neither is it to be adopted mindlessly. It’s a human tool. Creatives in all fields must therefore learn how to use it without surrendering their own humanity. If we become more productive and efficient in the mean-time, so much the better.