Add Category
As he attempts to tackle the New South Wales housing crisis, Planning Minister Paul Scully has been told he has "zero per cent" chance of meeting supply unless the government and sector immediately adopt AI technology.
Key points:
- The average housing project in NSW takes from 4 to 6 years for completion, which industry says is too slow
- Forecasts show NSW is 134,000 dwellings short of the population-based target
- Experts say AI tools for risk mitigation, design generation and rendering could save time and unlock more capital
Since taking on the portfolio Mr Scully has revealed the state is 134,000 dwellings short of the population-based target under the National Housing Accord by 2029.
That was before the recent increase in Accord numbers to 1.2 million, taking NSW's new target to 75,000 new homes a year.
Last week, at a Property Council event in Wollongong with the minister in attendance, industry leaders were asked to explore the expanding role of artificial intelligence (AI) in speeding up building approvals.
One engineer warned that the average project time in NSW of between four to six years from inception to completion was too slow.
Steve Cassells, engineering lead for engineering technology firm Neuron the state risks not being able to meet its housing targets unless it implements new processes.
"We have an almost zero per-cent chance of meeting the supply with our current process and timelines," Mr Cassells said.
Mr Cassells said without adopting the new technology now, the consequences for community would be "severe".
"We have major housing issues, we have well over a million homes we have to deliver in the next six years to keep up with demand," he said.
AI tech to cut red tape
Mr Cassells said existing AI technology was already available to address risk mitigation, design generation and rendering.
Importantly, he said those tools could be used to provide "early feedback on proposed developments" and help avoid months of wasted time and money and even refusals.
Mr Cassells said if councils supported the use of the tools, they could provide early in-principle feedback.
"Then a design could be made — and I use this word carefully — could be made 'approvable'," he said.
"Developers would get more certainty, which will unlock more capital to spend on refining concept designs knowing that these improved concept designs will get them a step closer to a DA approval."
The next priority use he said would need to be in "design optimisation" with systems to created to provide rapid feedback on utilities, wind acoustics, structure and visual impact for project approvals.
AI could streamline process
Mr Scully said he was "working on options" to improve the development assessment system.
"I've asked the department to scour the globe for the best examples of AI machine-learning technology in planning systems."
He said he was aware Seattle, Singapore and Brisbane had already integrated AI into their planning systems.
He told the conference he saw a role for AI in incorporating up-to-date data in transport assets, "considering the impact of changes on economic development on the way people engage with spaces" and more accurately assessing health and education needs of populations.
"Imagine running a proposal through a system to identify where there might be a problem and addressing if before submission.
"Or could AI be used to produce a better design that will reduce and limit community opposition."
He said it was important that AI was focused on where it was "appropriate and that its application remained transparent".
Gino Cavallaro, executive director of Digital Transformation from the NSW Department of Planning and Environment said enhancements had already been made to link development types and document requirements.
Mr Cavallaro said the department was also developing pilot programs to better inform development applications prior to lodgement.
"AI should not make a decision that a human would make, but aid and assist the process," he said.