AI has changed what professional proofreading looks like. Here's why that's good news for your publications budget.

It's a fair question, and one I've been asked more often lately. AI writing tools have come a long way, and if you've been using them for your communications, you've probably noticed that the results are – on the surface, at least – rather good.
So as a professional proofreader, am I worried? Not really. But not for the reasons you might expect.
What AI is actually getting right
Let's be honest: AI has raised the baseline. Content that used to arrive on my desk riddled with grammatical inconsistencies, erratic capitalisation and stray apostrophes now looks, on the surface, like polished copy. For anyone doing a quick visual scan, it can seem publish-ready.
That matters. It means the dreary, mechanical end of proofreading – fixing the same comma splice for the third time, querying a missing full stop – takes up less of my day, and that's not a bad thing.
What AI is quietly getting wrong
Here's where it gets interesting. AI tools are trained predominantly on American English, and it shows. I spend a notable portion of my time undoing:
- Em dashes used as punctuation – a very American habit, and one that jars in formal UK publications
- Serial (Oxford) commas before the final item in a list – grammatically defensible, but not standard UK style
- Smoothing out the AI voice – that oddly emphatic, slightly breathless quality that creeps in ('It is crucial that…', 'In today's fast-paced world…')
- Correcting misused words and phrases – AI is confident, but confidence is not the same as accuracy. It will reach for a plausible-sounding phrase and get it subtly wrong.
Will UK style simply drift towards American conventions over time? Possibly. I find myself wondering about that more often than I used to. But for now, organisations with a clear house style – particularly in the charity, education, faith and heritage sectors – rightly expect their publications to reflect it.
Where a professional proofreader earns their place
If AI is handling the surface-level tidying, it frees me to do something far more valuable.
The real expertise in editorial work was never about hunting for missing apostrophes. It was always about this:
Consistency. Does your organisation refer to itself the same way throughout? Are headings capitalised on the same principle in every section? Do your job titles, programme names, and place names follow a consistent pattern?
Usability. Can a reader navigate your document? Are contents pages accurate? Do cross-references actually lead somewhere? Are numbered lists in the right order?
Readability. Is this genuinely clear to your intended audience, or is it clear only to the person who wrote it? Are there sentences that make sense grammatically but would stop a reader cold?
Practical design integration.
Does the text fit the layout? Are there widows and orphans that need resolving? Do captions match images? Has a line break turned a heading into something unintentionally comic?
These are things that require human judgement, knowledge of your organisation and an understanding of how publications actually work.
The budget argument
There's a practical point here too. If you're commissioning professional editorial support, you want to know your budget is working hard. Paying a qualified proofreader to correct punctuation that AI could have caught first is not great value for anyone. But paying a qualified proofreader to ensure your annual report, fundraising appeal or school prospectus is genuinely consistent, navigable and fit for purpose? That's exactly where the investment pays off.
AI has changed what editorial support looks like. It hasn't changed whether you need it.
------
Kelly Owen is a CIEP Professional Member offering proofreading, editing and publications coordination to charities, faith organisations, independent schools, and heritage bodies. Find out more about me at
ultimateproof.co.uk




