Environmental impact assessment can be automated in digital planning

Out-Law News | 21 Sep 2021 | 12:09 pm | 3 min. read

The process of assessing the environmental impacts of proposed new developments can be automated, and the assessment itself “transformed into a real-time platform of information flows, analytics and predictive modelling”, a specialist in digital planning has said.

Dr Sue Chadwick of Pinsent Masons, the law firm behind Out-Law, said the human functions involved in completing environmental impact assessments (EIAs) were among several “administrative elements of the planning application process” that could be automated.

While Chadwick highlighted ongoing work by some local authorities in exploring the potential for administrative functions to be automated – including London Borough of Lambeth’s testing of artificial intelligence (AI) in the processing of planning applications – she said a shift to truly digital planning would be unlikely without changes being made to the regulatory framework.

Chadwick Sue_November 2019

Dr Sue Chadwick

Strategic Planning Advisor

If the planning system is to make the most of AI while still maintaining public trust, existing regulatory frameworks need to readjust and expand to accommodate its strengths and weaknesses

Chadwick’s comments were set out in a new report she authored for the Open Data Institute (ODI), on digital planning and its implications.

“There is significant scope to automate the way planning information is accessed and assessed,” Chadwick said. “Algorithms are much more efficient than humans in processing large and complex datasets; automation can also provide a rich, real-time record of environmental impacts and make sophisticated predictions about the impacts of development proposals with reduced investment of human time and resources. However, the complexity of the legal test regulating the decision-making function in planning makes it difficult to automate.”

In her report, Chadwick particularly highlighted the EIAs process as being “overripe for change”. She said the current “paper-heavy and impenetrable” process “involves submission of a daunting amount of largely static information which is then assessed by the local authority, involving the use of specialist advisers” while, as a report by the Digital Catapult notes, outputs from the assessments are “archived” and not “made available for re-use”.

The prospect of automating the EIAs process comes as developers are set to face new requirements around biodiversity arising from the new Environment Bill, which is currently progressing through the UK parliament.

Under the proposed reforms, developers seeking planning permission will be required to demonstrate a 10% biodiversity net gain from their plans based on the biodiversity metric the Department for Environment, Food and Rural Affairs (Defra) has developed.

While Chadwick highlighted the potential for algorithms to carry out “routine procedures or empirical assessments that do not need any human involvement”, she described the prospect of entirely machine-made planning decisions as “both dystopian and unrealistic”.

Chadwick said that there are ethical issues to consider around the shift to a digital planning system. The “core challenge” in this regard, she said, “is how to embed emerging principles of digital ethics as a complementary element in a system of law and policy that has yet to acknowledge that land has a digital identity”.

Although Chadwick used her report to flag the potential for digitisation of planning, she acknowledged the limitations of machines in the context of making planning decisions – from their lack of accountability and empathy, to their inability to “evaluate aesthetic merits as a human might”. She also highlighted the cultural barriers that need to be overcome for society to get “comfortable” with machine-made or machine-assisted decisions and identified the further legal barriers to the involvement of AI in planning decision making.

According to Chadwick, the core barrier preventing the move to digital planning is a legal framework for planning decisions that rests on principles established in 1946 when only Alan Turing could see a time when machines might take on some human functions. So long as the current system assumes that all decisions are “fully human”, there is potential for planning decisions made in full or in part reliance on algorithms, to be open to potential challenge, she said.

“If the planning system is to make the most of AI while still maintaining public trust, existing regulatory frameworks need to readjust and expand to accommodate its strengths and weaknesses,” Chadwick said.

The digitisation of the planning system in England was advocated by the UK government in its planning white paper published in 2020.

“The planning system must build on this success and follow other sectors in harnessing the benefits which digitisation can bring – real time information, high quality virtual simulation, straightforward end-to-end processes,” the government said in the white paper. “It should be based on data, not documents, inclusive for all members of society, and stimulate the innovation of the great British design industry.”

The reforms proposed in the white paper are expected to be introduced in the form of a new Planning Bill, which is expected to be tabled before parliament in the current parliamentary session.