AI Planning Objections: Why They Fail and What Actually Works

The Rise of AI-Generated Planning Objections

Over the past two years, a new category of planning representation has emerged on council portals across the UK. Letters that are lengthy, grammatically polished, and dense with apparent policy references — but that on closer inspection reveal a troubling pattern. The case law cited does not exist. The policies quoted are from the wrong local authority. The site-specific analysis bears no relationship to the actual proposal.

These are AI-generated planning objections, and they are now appearing in significant volumes. Research published in early 2026 found that nearly nine in ten local planning authorities reported receiving representations they believed to be AI-authored, with the majority saying the problem is growing. In one widely reported case, over five thousand objections were submitted against a single scheme — many near-identical, many raising arguments that had no connection to the site or the application before the council.

Several commercial services have emerged that offer to generate objection letters using large language models for fees as low as £45. The pitch is appealing: a professional-sounding letter in minutes, at a fraction of the cost of instructing a qualified planner. But the product these services deliver is fundamentally different from a professionally researched objection — and the consequences of relying on one can be serious.

The Core Problem: AI Cannot Research Your Case

A planning objection is not a piece of persuasive writing. It is a technical document that must do something very specific: demonstrate that a proposed development conflicts with the policies of the adopted development plan and would cause material planning harm. Doing this requires research that AI tools are simply not equipped to carry out.

The result is a letter that reads well on the surface but collapses under any professional scrutiny. Planning officers recognise these submissions immediately. They add to the officer’s workload — because every representation must be read and assessed — but they contribute nothing to the material planning analysis of the case.

The Hallucination Problem

AI language models do not understand truth. They predict the next plausible word in a sequence based on patterns in their training data. When asked to cite planning policy or case law, they produce text that looks like a real reference but may be entirely fabricated.

We have reviewed AI-generated objection letters submitted to councils where we were also acting on the same case. The pattern is consistent:

• Invented policy references — letters citing Local Plan policies by number that do not appear in the council’s adopted plan
• Fabricated appeal decisions — references to inspector decisions with plausible-sounding reference numbers (APP/X1234/W/24/...) that return no results on the Planning Inspectorate’s database
• Misattributed case law — court judgments cited with the wrong parties, wrong dates, or principles that the judgment did not establish
• Generic site descriptions — descriptions of the site and its surroundings that could apply to almost any residential street in the country, because the AI has never visited the site or viewed it on a map
• Wrong planning framework entirely — letters submitted for Welsh applications citing the NPPF (which applies only in England), or English applications referencing NPF4 (which applies only in Scotland)

A planning officer who identifies fabricated references will not simply ignore them. The letter’s credibility is destroyed. Every argument in it — including any that might have been valid — is tainted by the false material. The objector has spent money and time on a document that actively undermines their case.

The Planning Inspectorate’s Response

The Planning Inspectorate (PINS) has taken a clear position on the use of AI in planning casework. In guidance published on GOV.UK, PINS now requires that any party using AI to create or substantially change any part of a submission must disclose this. The disclosure must identify the AI tool used, what it was used for, and what verification steps were taken to ensure accuracy.

The guidance contains a direct warning: the Inspectorate states that improper use of AI could be treated as unreasonable behaviour — and that parties who submit AI-generated material that adds unnecessary burden to a case are at risk of a costs award.

The Inspectorate’s position is clear: AI can be used as a tool to assist with formatting, translation, or accessibility. But the person submitting the evidence bears full responsibility for its accuracy. The “golden rule,” as PINS describes it, is that you must use AI responsibly and ensure that everything it generates is accurate and appropriate.

The Orkney Incident: When AI Crosses a Line

In one of the most striking examples to date, Orkney Islands Council launched an investigation after fake AI-generated objection letters were submitted against a proposed hotel development in Kirkwall. The letters used the names and addresses of real local residents and businesses — without their knowledge or consent. The developer recognised the impersonated individuals personally and alerted the council.

The council removed the fraudulent submissions from its planning portal and opened an investigation. The incident is believed to be among the first of its kind since the introduction of digital planning portals, and it raised serious questions about identity verification, the integrity of online consultation, and the potential for AI tools to be weaponised against legitimate development.

While this is an extreme case, it illustrates a broader point: AI-generated objections do not just risk being ineffective. They risk actively harming the planning process and the people who depend on it.

Why Councils Are Not Fooled

Planning officers read objection letters for a living. They are trained to assess representations against the development plan and to identify which arguments raise genuine material planning considerations. AI-generated letters share characteristics that are immediately recognisable to an experienced officer:

• Polished but generic. The language is fluent and grammatically perfect, but contains no site-specific detail, no reference to particular windows, boundaries, or streets, and no evidence of having read the submitted plans
• Verbose and repetitive. AI tends to produce lengthy text that restates the same point in multiple ways. A professional objection is concise and precise — every sentence earns its place
• Scattergun approach. Rather than identifying the two or three strongest grounds and building a case around them, AI letters throw every possible objection at the wall — including non-material matters like property value and loss of view — hoping something sticks
• Identical across multiple objectors. When several residents use the same AI tool on the same application, the resulting letters are near-identical. Officers cluster them as a single representation, not multiple independent objections

The irony is that a shorter, rougher, genuinely personal letter from a neighbour who describes the actual impact on their actual home carries more weight than a five-page AI essay that could have been written about any development anywhere.

What Actually Works: The Professional Alternative

An effective planning objection is the product of research, professional knowledge, and site-specific analysis. It cannot be generated in minutes because the work that underpins it — reading the application, identifying the policy conflicts, assessing the site context, researching comparable decisions — takes time and expertise.

When Planning Voice prepares an objection letter, the process involves:

1. Full review of the application. We access the council’s planning portal and read every submitted document — plans, elevations, design and access statement, technical reports, and any supporting material. We understand what is actually being proposed before we write a single word.
2. Local Plan policy research. We identify the specific policies in your council’s adopted Local Plan that the proposal engages, and assess whether the scheme complies or conflicts. This is the foundation of every argument we make.
3. National policy analysis. We apply the relevant provisions of the NPPF 2024 (in England), NPF4 (in Scotland), Planning Policy Wales, or the SPPS (in Northern Ireland) to the facts of your case.
4. Site-specific harm assessment. We analyse how the development would affect your property and your neighbourhood — light, privacy, character, traffic, noise — using established methodologies including BRE daylight guidance, separation distance standards, and local design codes.
5. Comparable decisions. Where relevant, we identify real appeal decisions where planning inspectors have refused similar schemes on similar grounds. Every reference we cite is verifiable.
6. Bespoke drafting. Every letter is written from scratch for your specific case. No templates. No AI-generated text. A qualified Chartered Town Planner researches and writes your objection personally.

The result is a letter that the planning officer must engage with substantively — because it speaks their professional language, references the correct policies, and demonstrates specific, evidence-based harm.

AI as a Tool, Not a Planner

None of this means AI has no role in planning. AI tools can be genuinely useful for tasks like translating documents, improving the accessibility of written material, summarising lengthy technical reports, or helping people articulate concerns they struggle to put into words. The Planning Inspectorate recognises these legitimate uses.

The problem arises when AI is asked to do the planner’s job — to research policy, assess harm, cite case law, and produce a technical document that will be relied upon in a quasi-judicial decision-making process. That is professional work that requires professional judgement, accountability, and accuracy. An AI model has none of these things.

If you use AI to help draft your personal comments on an application — to organise your thoughts or improve your phrasing — that is a reasonable use, provided the underlying concerns are your own and the facts are accurate. But if you are relying on an AI service to produce a standalone objection letter that you submit as your representation, you should understand what you are actually getting: a document that looks professional but has not been professionally prepared.

The Bottom Line

Planning objections succeed or fail on the strength of their planning arguments — not on how polished the prose sounds. A well-researched letter grounded in the correct Local Plan policies, supported by real evidence and genuine case law, written by someone who has actually read the application and understands the site, will always outperform an AI-generated letter that has done none of these things.

If your home, your neighbourhood, or your quality of life is at stake, the question is not whether you can get a cheaper letter from an AI tool. The question is whether you can afford to submit one that does not work.