AI as a research assistant for planners

While general AI models, which draw from broad, often unrelated, data sources can be very useful, significantly greater value can be derived from models which are able to focus on user-specific documents and datasets – this capability has been improving notably over the last year. 

Such tailored models could act as powerful digital research assistants that are designed to help planners navigate large volumes of information efficiently. These tools offer the ability to upload policy documents, planning applications, environmental impact assessments, research papers, and even web links to planning committee meetings. The tailored AI tools could then analyse and summarise the material, utilising their ability to understand context and relationships between different pieces of information, and also respond to specific queries from the user, such as requesting a BNG figure that may be hidden within a dense report.

For example, a planner could ask the AI tool to compare emerging Local Plan policies with the National Planning Policy Framework to identify potential conflicts. AI systems could also cross-reference new applications against a database of appeal decisions, helping officers identify potential issues early on, and providing insights based on precedent. This can lead to more consistent decision-making and reduce the likelihood of costly appeals. This functionality effectively creates a centralised, searchable repository, enhancing policy analysis and helping to ensure better consistency in planning decisions. It also creates a more interactive and engaging research experience through natural language conversations, much like consulting a colleague who is an expert on a particular topic.

Importantly, the majority of these emerging tailored AI tools offer features like citations, which link responses directly back to the source material. This allows planners to verify information easily, examine specific details, and promotes transparency in AI outputs. For instance, if a planner asks about the environmental impacts of a proposed development, an AI tool could provide a summary drawn from uploaded environmental impact assessments, complete with citations to the exact sections of the document. This not only saves time but also ensures that critical details are not overlooked, ultimately supporting more informed decision-making.

AI planning experts 

Creating AI "experts" tailored to specific sets of planning-related material opens up a range of potential use cases in planning. For instance, AI systems could become specialists in Local Plan or Development Consent Order (DCO) examination libraries, assisting planners, inspectors, and stakeholders by quickly retrieving relevant policies, past decisions, and evidence. Given the rapid pace of advancements in AI, it may soon be possible to extend this to decision-making as well, for instance, through the determination of straightforward permitted development applications where regulations are often unambiguous, further streamlining the planning process.

There is also potential to integrate an AI layer onto local planning authority websites to help officers, councillors, and members of the public better understand the content of submitted planning applications, and the feedback they receive. Planning documents are often dense and filled with technical jargon, making it challenging for the public to grasp the implications of proposed developments. An AI assistant, via a chat interface, or potentially via voice in the future, could simplify these documents by providing clear, accessible summaries, and answering questions in everyday language, and in multiple languages.

Challenges and considerations

While AI offers significant potential for transforming the planning system, it is important for everyone to approach these tools critically. AI models can sometimes produce inaccurate or biased information, which was highlighted in a recent Ministry of Housing, Communities and Local Government study on using AI to streamline the planning process. The study found the accuracy of AI responses ranged from 43% to 76%. The ability to provide information sources noted above is an important step in combatting this issue, but human oversight by planning professionals remains crucial. Planners need training in the use and limitations of AI systems, which allows them to apply their expertise and critical thinking to verify AI-generated outputs and ensure accuracy and reliability in their work.

The Planning Inspectorate (PINS) has recently published guidance on how AI should be "responsibly and lawfully" used in planning, emphasising that planners must disclose if AI has been used to create or alter any documents, including providing details about the AI tools and data sources. This kind of transparency is crucial to maintaining trust in the planning process and ensuring that AI is used responsibly.

Another challenge lies in integrating AI tools into existing systems. Planning departments often use legacy software and fragmented data systems, making seamless integration a technical hurdle. Investment in infrastructure, establishing consistent data standards, and collaboration between IT specialists and planning professionals will be key to overcoming these obstacles. Central to this is ensuring data privacy and security, as planning documents often contain sensitive information, and thus robust security measures in IT infrastructure are essential to protect data from unauthorised access or breaches, maintaining the integrity of the planning process.

Conclusion

The integration of AI into the UK planning system could offer a significant step towards greater efficiency, transparency, and responsiveness. For organisations, such as local authorities, it is essential to have a well-defined technology strategy in place to ensure AI complements and enhances existing functions, rather than merely serving as an add-on. 

By creating AI systems capable of managing vast amounts of information, enhancing decision-making processes, and improving communication among stakeholders, the potential for meaningful transformation in the planning system is significant. To fully harness this potential, collaboration between policymakers, technology developers, and planning professionals will be necessary to ensure that AI tools are developed and utilised in ways that truly benefit the planning system.

Further information

Contact Ciaran Hagan