Why You Still Need a Human Editor Part 3: About AI & Written Language

I’ve been a science fiction reader for much of my life, and I’m fascinated by the possibilities of artificial intelligence. However, as a professional wordsmith, I have reasons to be concerned about our society’s current love affair with LLMs. I don’t believe that human-generated writing will (or can) ever be fully replicated by AI tools – certainly not the material that people read for enjoyment. But I recognize that AI-generated business content is growing in popularity because it saves effort, time, and money. That honeymoon will end at some point. Not holding my breath for it, though.

Photo by Feodor Chistyakov on Unsplash

The Sound of Our Words

 I was already learning about what AI could and couldn’t do a few years before ChatGPT, Claude, and MS Copilot became household names. Clients were giving me interview or podcast transcripts as source material, and it eventually became more cost-effective for them to run the audio through software rather than hire a human transcriber. I came to understand that neither human nor AI is perfect, but the quality of error differs from one to the other.

• Humans understand how to show when the speaker pauses, stammers, or restarts a sentence. AI, not so much.

• Humans might have a better understanding of the speaker's idioms, turns of phrase, and how a regional dialect or a non-English-language accent can make certain words sound completely different. AI? Again, not so much.

• On the other hand, AI works faster, cheaper, and doesn't have bad days expressed via laziness or Freudian slips (such as the proper name “Judy Drury” rendered as the looming obligation of “jury duty”).

• Specialized language can be a downside when working from a transcript. If the speaker is naming names, using specialized terms, or dropping lots of acronyms, both human and AI get confused. (This is a situation that might call for a technical edit, as described in Part 1.)

Speaking as a Writer

 Some people are starry-eyed about the sea change that LLMs are bringing to how we communicate. Other people are calling this inflection point the end of the world as we know it. And copywriters (e.g. those who write nonfictional/operational/instructional prose for money) are caught in the middle of this swirl of love and hate for AI.

Spend an hour reading conversations on LinkedIn, and you’ll find two flavors of heavy expectation:

1) “If you want to stay relevant (and employed), make AI work for you. Follow these vague suggestions (so obvious to me that I won’t elaborate), use your brain, and figure out some angle that works for you.”

2) “Time will tell if this whole AI business is a fad or an abomination. But starting here and now, the only thing any self-respecting writer can do is fight back by going big with vaguely defined human qualities that no machine can replicate (so obvious to me that I won’t elaborate).”

As a writer, I love writing, and nothing can change that. As a professional content creator, I feel like there’s a struggle ahead. AI might not doom the content/editorial sector, but it will change it. While highly creative and stylized genres are less likely to be affected, the main impact will be on business writing: web and marketing content, white papers, presentation decks, etc. I’ve heard predictions that writers won’t be replaced so much as expected to add AI tools to the writing process – and therefore expected to deliver a bigger workload.

I would conclude by saying that there will always be a human factor in play. No matter how well your AI generates text, if you’re looking for a specific reaction from humans, you’re still going to need human eyes on your final text before you post or hit send.

This has been the final blog of a three-part series about editing. In case you missed the others, please read:

• Part 1: The Craft of Editing

• Part 2: Issues That I Help Writers Avoid