I've been reflecting how, this time last year, most of my proofreading work looked the way it always had. I received a proof and checked it for typos, clunky sentences and consistency – something I've been doing quite happily for the past 25 years – but I've noticed it's definitely starting to change.

More of the documents I receive now started life somewhere other than my client's desk; some have been drafted in ChatGPT, others 'magically' designed in Canva, or text has had a quick 'tidy up' by Copilot before it reached me.

I've noticed that the types of errors I'm catching have changed too. I'm correcting punctuation that's technically correct but stylistically wrong for the client (serial commas anyone?). I'm smoothing out sentences that are grammatically flawless but somehow don't sound like anyone in particular. I'm also flagging issues with tone that's now oddly formal, because that's what AI tends to default to.

None of this is proofreading in the traditional sense; it's something closer to voice alignment – and it's becoming one of the most valuable things I do.

Why this matters more than it might seem

AI is genuinely useful; I've literally just finished landscaping my garden with the help of Copilot! I'm a language lover and prefer paper over screens any day of the week – but I'm not precious about where a first draft comes from – and I say so to clients regularly. However, AI doesn't know your organisation. It doesn't know that your charity's newsletter has always had a slightly wry, conversational tone, or that your school's prospectus needs to sound aspirational without being corporate. It produces content that's plausible – and plausible isn't the same as yours.

The risk isn't that AI-assisted content is obviously bad; it's that it's subtly, almost invisibly 'off' – sounding confident and correct but not quite right. It's something a reader can't always name but somehow notices – and as a proofreader whose radar is permanently 'on', I'm seeing this more and more.

What I'm doing differently

I've started naming this work explicitly as part of my proofreading and editing service: brand voice and consistency review, sitting alongside the traditional checks for spelling, grammar and formatting. In practice, it means I'm not just asking 'is this correct?' but 'does this sound like you?' – checking content against what your organisation has published before, your stated tone of voice and what your audience expects to hear from you.

For clients producing more content than ever, with more tools and more contributors involved, that consistency is becoming harder to hold onto without someone dedicated to watching for it.

Where I stand on AI

I don't run client content through generative AI tools myself – that review is always done by hand, and that's not changing. But I'm happy to work with content that started with AI, and I kind of expect that nowadays. In fact, I'd rather you draft quickly with whatever tool helps you, and let me make sure what goes out the door is accurate, consistent and recognisably yours.

Whether I like it or not, that, I think, is what proofreading is becoming for a lot of us – a partnership with AI, where the human part of the process is more important than ever.

If your team is producing more AI-assisted content and you're not quite sure it still sounds like you, I'd love to help. Get in touch or book a free call.