How is AI changing the planning process?
Digitisation in planning is a no-brainer. The huge amount of data across the built environment needs to be unlocked, with software made available without significant cost or technical barriers. We can already see the efficiencies emerging across the planning processes: GIS layering of constraints and opportunities; site selection for local plans; engagement in planning; visualisation of proposed development; automating validation and assessment of straightforward applications; and chatbots to answer questions. We can imagine many more benefits. As the scope for AI is debated across the industry, it is already being promoted by DLUHC through a number of sponsored pilots*.
The AI Transformation
The publication of ChatGPT and its contemporaries in the last 6 months has changed our horizons. They show us that algorithms and machine learning can go a long way – and that AI should be warmly welcomed for its transformational prospects, notwithstanding a current reputation of being ‘confidently wrong’. However, our eyes remain wide open to the risks it brings with it and there are some fundamental questions in relation to planning.
1 Context
We should question whether AI is capable of the broad understanding of multiple topics that play together in planning – now and in the future. Beyond technical issues like zoning, traffic patterns and land mapping, how will it interpret differing geographical contexts, changing political environments and social expectations? How will it anticipate economic cycles and avoid political bias?
2 Judgement and Creative Thinking
Applying planning policies flexibly when required, or weighing competing policies against one another, are a real test of machine learning. Could AI be capable of policy interpretation and intelligence, prioritisation between issues and appropriate application of case law? How will it handle the strings of conditions, exclusions and amenity judgements introduced by some more recent PD Rights, and come to a fair judgement? There is justified scepticism around critical and creative thinking, as to how all the relevant factors should be applied, in, for example, resolving contentious issues or the interpretation of ‘beauty’ in the built environment.
3 Human Understanding
Although AI is capable of a nuanced understanding of language, digesting and extracting the essential planning points from a ‘stream of consciousness’ response to a consultation is another challenge entirely. Could it accurately represent a highly considered detailed consultation response that doesn’t fit in the box?
4 The Trust Issue
Finally, and most importantly, will AI be able to discern reliable data and misinformation? Will it be aware of errors and risks and tell us? Unlike government, AI is not subject to data regulation. Indeed the scope to manipulate AI visualisations** of proposals and consultation responses already exists and is becoming increasingly sophisticated. Can we trust AI?
Mapping the future
If we want our built environment to reflect the world we want to live in, then we should tread cautiously. Doubtless, straightforward cases will be automated, but as with the medical and legal professions, we should reserve more complex decisions for human consideration and overview.
Through augmented tools and platforms, the speed and efficiency of AI are hugely beneficial to the planning process. Let’s just hope we’ll avoid a long trail of unintended, inappropriate developments resulting from an AI case officer saying ‘yes’.
Frances Wheat
References
*Lessons learnt from Round 1 pilots https://dluhcdigital.blog.gov.uk/2023/04/03/what-we-learned-from-round-1-proptech-innovation-fund-digital-citizen-engagement-pilots/ and https://www.localdigital.gov.uk/digital-planning/case-studies/
**Round 3 Improving visualisation and managing engagement data https://dluhcdigital.blog.gov.uk/2023/03/23/15-council-led-projects-funded-to-scale-digital-solutions-which-aim-to-improve-citizen-engagement-within-the-planning-process/
