Human or ChatGPT for proofreading? Six considerations

Mistakes are nothing to be afraid of.

Own them, deal with them, learn from them.

If it’s a particularly bad blunder, thank whoever pointed it out, and apologise with sincerity. Fix it if you can.

If it’s the entertaining kind of blunder, laugh about it! Sometimes writing mistakes are unintentionally hilarious.

No-one writes perfectly, and no-one catches all their own errors. Not even a professional proofreader.

But what a professional proofreader can do is catch all of yours.

So, why wouldn’t you enlist one to review your work?

Well, they cost money. And if you’re working to a deadline, involving someone else will just add extra time.

But you know what’s free, and provides almost instant feedback?

ChatGPT!¹

My advice: you shouldn’t rely on ChatGPT to proofread your work. But I would say that, wouldn’t I?

So, judge for yourself. Check out the six considerations below, all of which you should take into account before deciding whether you need a second pair of eyes on your work, or a second pair of AIs.²

1. Creativity

Human editors and proofreaders are extremely nerdy – and well-practised – when it comes to wording, style and structure. We live to tweak text, hit those perfect beats, and melt your words into the smoothest flow.

ChatGPT can certainly appear creative. And if you request text in a certain format, it can deliver. But the results are usually pretty generic, even cringey when it decides to go all over the shop with adjectives. It also tends to repeat patterns a lot. How do you think AI detectors work?

2. Context

Unlike Large Language Models (LLMs), humans understand concepts like context and cultural nuance. These are very much human constructs, after all.

This means a human (with the necessary experience) can keep your content accurate, natural sounding, and contextually appropriate.

Should you trust ChatGPT to contribute to work covering a sensitive topic? Well, even if mistakes are nothing to be afraid of, there are still some you may want to avoid.

3. Accountability

When you hire an editor or proofreader, you don’t just get the end product, i.e. a wonderfully written, punchy, engaging and typo-free piece of content.

You also get a faster way to learn from your mistakes.

Good editors and proofreaders will track all of their changes, and provide explanatory comments as needed. They will also be happy to chat further and help you understand particular edits.

And in the rare case of an incorrect edit, a good editor or proofreader will own their mistake, apologise sincerely, and fix it for you. Because we learn from our mistakes too, and because we do what ChatGPT doesn’t: we take accountability.

What if an editing error did result in a particularly bad blunder? If you used ChatGPT, you have no recourse. It’s all on you.

(Also, welcome to the scary side of my job. This is why we are so very careful.)

4. Accuracy

Continuing that thought, a human editor or proofreader is not going to introduce random rubbish into your content. Nor are we going to change the meaning of your words. We have reputations to uphold, and we care.

Have you heard about how ChatGPT ‘hallucinates’? This is the term used to describe an AI chatbot pulling completely made-up statements, false supporting evidence, and even fake links to that ‘evidence’, out of thin air.

I gave it a try myself once. I asked it about an art concept that I had made up on the spot. The amount of information I got back, including URLs that went nowhere, was astounding!

I also asked ChatGPT to proofread this article, into which I introduced a couple of deliberate typos. It missed both typos, and instead suggested I make changes to a sentence that I hadn’t written in the first place.

It’s risky to rely on ChatGPT to suggest or make edits to your work.

5. Freshness

We are all alive in the present moment. Whoa.

ChatGPT, on the other hand, lives only in the past. Its index only holds content up to a certain date.

If you’re writing about recent news, a new product or trend, or anything up-to-date really, it’s not going to be of much use. Too bad for ChatGPT, new stuff just keeps happening, all the time.

It’s not going to be great with the latest, of-the-moment content trends either, such as cutting-edge marketing copy, or text to fit a new social media format.

6. Confidentiality

A professional editor or proofreader will treat your work confidentially. They will sign an NDA if you ask them to, legally binding them to keep their mouth zipped.

ChatGPT learns from whatever you type into it. Then, it will proudly parrot back your confidential, proprietary content to other users.

Like your competitors.

Is ChatGPT really so bad?

I don’t hate ChatGPT. It’s great for knocking ideas around, and when it does generate text, it’s usually typo-free. I mostly use it to help build my word hoard when I’m starting a creative writing piece.

But when it comes to reviewing your work, I would always recommend working with a human. Until the machines finally take over, this is the only way to ensure your content is dealt with professionally, confidentially, and to a high standard.