Short answer: technically yes.
Long answer: absolutely not, unless you fancy a personal introduction to your local enforcement officer.
Let’s Start With the Obvious
You’ve probably seen it online – someone bragging that they used ‘AI’ to write their planning application. They typed, ‘Write a planning statement for a dog field in West Sussex’, and got a shiny, polite essay about ‘enhancing canine wellbeing’ and ‘improving rural amenity’. It even had subheadings.
Looked official. Sounded smart.
Got refused in eight weeks flat.
AI can generate words. It cannot understand context. And planning, more than almost anything, is 90% context.
Here’s What ChatGPT Doesn’t Know
Your site.
It doesn’t know your access visibility splays, whether the lane floods, or if the neighbour’s cat is already plotting against you.
Your council.
It doesn’t know that East Suffolk adores ‘biodiversity enhancement’ but loathes ‘change of use’, or that one Devon officer considers bark noise a ‘significant environmental impact’.
Your policy environment.
The GPDO, NPPF, and every Local Plan in England change constantly. AI is trained on old text – not the version that matters today.
Tone.
Planning officers smell fake confidence a mile off. They want clarity, not waffle. AI writes like a student who’s just discovered thesaurus.com.
Of course – you can inform AI of all these things but honestly, by the time you have assimilated everything it needs to make a marginally better stab at it, you might as well have written it yourself.
We’ve Tested It (So You Don’t Have To)
Because we’re gluttons for punishment, we’ve run a few experiments. We asked ChatGPT to write a Design and Access Statement for a real dog field site (names changed to protect the guilty).
It wrote 1,200 words of lyrical nonsense about ‘community spaces’, ‘urban dog parks’, and ‘the sustainable social fabric of pet ownership’. Lovely, except the site was in rural Devon, population 400, and the only ‘community space’ nearby is the pub car park.
It then confidently misquoted planning policy that doesn’t even apply in that district.
And it missed the single biggest risk on the site – highway visibility.
A human consultant reads a map. ChatGPT just guesses.
The Real Danger Isn’t What It Gets Wrong – It’s What It Gets Nearly Right
This is the tricky bit.
An AI-generated statement looks convincing. It’s grammatically perfect, full of buzzwords, and long enough to feel expensive.
But it’s built on sand.
A planning officer reading it might not even spot the holes immediately – they’ll just send a ‘Request for Further Information’, which triggers a three-week delay. Then another. And then your application and enthusiasm times out.
Suddenly, you’ve wasted months and your ‘AI efficiency’ has turned into a slow-motion disaster.
Could AI Ever Actually Do It?
No doubt.
If machines keep improving at the rate they are, it’s perfectly possible that an AI Agent will eventually draft, check, and even submit planning applications. It’ll probably even remember to include the right visibility splay diagrams – though whether it can charm a local planning officer over Zoom remains to be seen.
Right now, the cleverest use of AI is as a research sidekick. It can help you dig through policy, summarise documents, or turn technical jargon into something you can actually understand. But it only works if you already know what you’re looking for. Otherwise, it’s just confidently inventing nonsense with good punctuation.
When we built our own AI planning experiment – generating a full Design and Access Statement using several different tools – it wasn’t because we fancied the easy route. It was because we wanted to know whether we were about to have an awkward conversation with our long-standing partners, Evolve Planning and Design, where we essentially fired them.
Spoiler: Evolve are still very much employed.
The AI statement looked fine – polished, wordy, and full of ‘enhanced amenity provision’. But when a proper consultant reviewed it, they found errors we hadn’t even noticed. It had all the right vibes of professionalism, and none of the substance.
AI might get there one day. Until then, it’s that intern who turns up in shiny shoes, talks a big game, and has no idea what ‘context’ actually means.
It’s Not That AI Can’t Help – It’s Just That It Shouldn’t Lead
There is a place for ChatGPT in the process – and it’s not in writing your planning statement from scratch.
It can help you: brainstorm your project summary, draft a polite letter to your parish council, or simplify an officer’s response so you can actually understand what they’re saying.
But it should never, ever be the author of your application. It’s a tool, not a substitute for professional competence.
What Happens When AI Meets Enforcement
We’ve seen it happen. Someone uploads an AI-written statement, gets refused, then pays for professional help anyway – along with the added embarrassment of the officer referring to their ‘automatically generated’ submission in the decision notice.
It’s like trying to home-brew your own penicillin because you read about it on Reddit.
What to Use AI For
- Use AI for research and preparation – not decision-making.
- Use it to understand terms, find examples, or check your spelling.
- Making videos of your dogs riding bicycles, wearing top hats and tinsel scarves
Then, once you think you’ve cracked it, run it past someone who’s actually dealt with your local authority before.
The difference between ‘looks plausible’ and ‘gets approved’ is about 20 years of pattern recognition – and no machine can replicate that (yet).
Final Word
ChatGPT can write a planning application.
It just can’t get it approved.
If you want something that reads beautifully but ends up filed neatly under ‘nice try’, sure – let the robots have a go.
If you’d like a field that’s actually open and earning, speak to someone who knows where the traps are buried.👉 To see how we combine data, experience, and actual humans to get applications through, read about our partnership with Evolve Planning and Design.