Only a couple of topics ignite the communities more violently than what to do with land. The prospect of publishing small parts of Green Belt Land for residential developments can transform a wind farm proposal or plans for a brand new road into mildly man-made residents Passionate supporters overnight.
This visceral connection between humans and the situation perfectly shows the famous commentary that “All politics is local”. In England is the principle that each citizen needs to be given the chance to discuss planning matters legally anchored. Before a planning document is accepted, local authorities must give the general public the chance to provide feedback.
The logic for this is predicated on a standard sense morality: Before binding decisions are made about how an area could change, the locals who must live with these decisions needs to be given the chance to support or reject this plan.
In practice, that is an especially cumbersome process. The local authorities have to grasp 1000’s of comments. This prompted my colleagues and me on the University of Liverpool to take into consideration how AI might be used to make this process more efficient.
As soon as an area authority publishes the corresponding local planning document, every citizen, company, public, private or third-party firms has the correct to submit a written answer. These can take care of your complete document or consider a selected problem.
In all cases, the local authority is obliged to record, understand and mix all public entries. You then resolve whether the document requires changes or whether further evidence is required to justify the suggestions.
This creates an amazing burden on planning departments within the country. In areas with high development, the submissions often count within the tens of 1000’s. And individual submissions range from some sets to over 100 pages.
Planner must read, absorb and synthesize all this information in a final report that’s used for decision. This report must adequately represent the aggregated views in all submissions.
In addition to the mere volume of the reactions, human cognitive restrictions and distortions make the method difficult. Some submissions might be emphasized greater than others. Recently read submissions probably have a greater influence on the reader than before.
A digital solution
These challenges prompted us to explore alternatives. We have brought along with the Greater Cambridge Shared planning for Cambridge City and South Cambridgeshire District Councils to develop a AI-driven solution. Our tool, Plan AI, would read and summarize public submissions on the planning process.
In 2025 my colleagues and I carried out an actual experiment. Three live public advisory exercises were processed in parallel – once by planners and once by Plan AI.
Overall, a planning officer took 320 submissions to download and process. Eighteen hours of this time were used to summarize each submission – a task that the AI only lasted 16 minutes. During this time, the AI tool was also capable of create comprehensive reports during which essential topics were identified, the sources and the geographical evaluation of the submissions were identified.
A subsequent qualitative evaluation showed that the standard of the summaries generated by the human planning officer and the summaries created by Plan AI doesn’t give any apparent difference. In fact, the final overview document created by Plan AI is a big addition to what would normally be produced. It contained a geographical evaluation of the origins of submissions – crucial information for planners to grasp which communities and demographic groups took part within the consultation.
PJHPix/Shutterstock
The way forward for planning
The British government has determined A vision So that the local authorities can take AI to scale back the executive burden and to enhance the federal government's efficiency. For example recently Ruck out a AI toolDeveloped with Google Deepmind to digitize planning documents.
The effects of experiments like these are far -reaching. Planners can consider their core competence – the assessment of applications and the support of the federal government of the federal government for housing, latest cities and renewal of the infrastructure – as a substitute of spending countless hours to process public comments.
AI can process enormous amounts of text consistent and comprehensive than humans. It also can discover connections between submissions which will otherwise be ignored.
In view of the executive burden, the local authorities could possibly seek the advice of residents more often in a broader area of planning issues, which makes planning much more democratic. Planners that were free of paperwork could also devote more time to a meaningful public commitment.
One danger with AI is in fact that it might be used on the opposite side of the consultation to create a big submission volume so as to consider a certain perspective. However, AI tools might be used to defend themselves.
Planai or similar programs can create a right away summary of a comment in, a super opportunity to insert a review of the review, which the input is definitely human. If man is utilized in this manner within the loop, the potential that the AI is used shall be used to distort consultations.
By increase the correct tools and systems, we are able to create planning processes that react each more efficiently and to the influence of the citizen – a victory for democracy and effective government equally.
