The budget promise of £65 million represents around £150,000 per planning authority. Welcomed though I’m sure any additional funds will be, it may not strike many as being a game changing amount of money.

Aimed at reaping the rewards of digitising the planning system, what could this achieve? Implementing some blue-sky thinking, £65m could actually result in significant changes to the system.

What if planning applications could be determined by machine learning? Machine learning is a process whereby an artificial intelligence is trained by being fed a series of known data inputs and their resulting outcomes. The artificial intelligence ‘learns’ the key factors behind the decision outcome. This then enables the algorithm to be fed a new set of inputs and use its ‘learning’ to determine the outcome. The database of inputs is readily available on planning authorities’ planning portals, potentially providing the ‘big data’ that machine learning needs. They also contain the binary outcomes: approved or rejected.

Once trained, machine learning could be used to validate applications and automatically inform applicants of defects in applications, cleverly and clearly recommending the steps needed to secure validation. 

Machine learning could also quickly identify those types of application that need special treatment, such as EIA development, or applications that need referring to the Secretary of State. In addition, it could be used to deal with the administratively heavy statutory requirements of the notification and publication of applications. 

Going one step further (perhaps too far?), machine learning could even determine the planning applications themselves, and then approve the discharge of conditions too.

In each case the determination could be so fast that you could ask for a ‘pre-determination’ and, akin to checking your credit score online, the system could politely inform you that your application has zero prospects of success, but if you still want to progress and pay the application fee you could still do so.

Of course, allowing a computer to make planning decisions would be quite controversial (at least initially, until we are all comfortable with the idea). The machine learning output could stop short of making the decision and simply provide a recommendation, allowing the planning officer to have the final say on the planning application. The appeal system also has a huge database of historic decisions that could lend itself to machine learning.

A pilot would be needed and this would consume a big chunk of the funding, but the potentially huge gains in resourcing the determination of planning applications would make this so worthwhile. 

If not machine learning, then how else could the system benefit from further digitisation? 

The example given by the government of checking the scale and north arrow on a plan is a concrete one, but with marginal gains in efficiency. Likewise, digitised plans could be digitally stamped ‘approved’ and permission drafting could no doubt be automated too. Planning registers would also likely have the potential for digitisation gains.

Planning authorities are already banding together to share resources. The cost-reduction potential of realising digitisation gains might accelerate this process.

At the Built Environment Committee meeting in early November, the right honourable Chris Pincher explained to the members that local plans ought to be capable of being produced in as little as 30 months. The minister regularly returned to the theme of digitisation and he summarily set out how this was key to achieving that local plan delivery timescale by removing day-to-day administration, streamlining the planning system and ‘increasing the bandwidth’ of local planning authorities to leave room for strategic decision making. However, he did not give the details of what that streamlining and administrative efficiency would entail. 

With the wider reshaping of the planning system as yet unclear following the debates in reaction to the “Planning for the Future” white paper, digitisation is a common sense next step. Investments in technology and digital tools will undoubtedly benefit the planning system – the question is now whether we choose to go for a small step forward by further deploying existing tools or a leap into the future with new machine learning capabilities.

This article was first published in The Planner

This publication is intended for general guidance and represents our understanding of the relevant law and practice as at November 2021. Specific advice should be sought for specific cases. For more information see our terms & conditions.


Date published

16 November 2021


View all