Dr. Xinyu Fu is an urban planning researcher at the University of Waikato. He’s currently researching the implications that Large Language Models (LLMs) such as OpenAI’s ChatGPT could have on the future of urban planning. His most recent research demonstrates how powerful AI tools can significantly reduce the time needed to analyze thousands of public submissions on planning proposals. Understandably, this task often requires a substantial investment of time and resources. By harnessing AI, Dr. Fu suggests that urban planners can devote their efforts to more critical aspects of their work.
Dr. Fu’s research demonstrates that LLMs can effectively condense piles of thousands of public submissions in no time. This efficiency would reduce unnecessary hours of manual labor and bring the cost down to under $US15. Beyond saving time, this efficiency has the potential to change the game for planners who today find themselves bogged down by labor-intensive administrative tasks. When speaking to a number of planners, Dr. Fu discovered that most were intimidated by these duties. This pressure made it extremely difficult for them to truly engage with communities and think long-term.
Australian Treasurer Jim Chalmers has recently pointed out the mounting economic headwinds. In reply, while the potential productivity benefits of increasingly using AI technologies in urban planning would be dramatic, they must be weighed against new risks to job security and a clear need for new regulations.
The Role of AI in Urban Planning
As an urban planner, Dr. Fu is adamant that AI should never replace the work done by human planners. Rather, it is a replacement for the sophisticated analytics that they are used to employing themselves. “They spend a lot of time on doing those [administrative tasks], and a lot of those tasks can be augmented by AI,” he stated. He mentioned that planners would be able to quickly and easily reproduce even the most complex tasks with LLMs. He continued, “We just played around with the AI, OpenAI’s GPT-4. We take that and simulate what planners are trying to accomplish.
The potential advantages extend beyond mere efficiency. Prior to LLMs, the tools available were extremely sparse, Dr. Fu told us. “Now, we can do in hours what used to take weeks or even months.” The potential of this capability to transform urban planning processes is profound. It inspires planners to focus more on engagement and understanding community needs than on the day-to-day minutiae of project delivery.
Ms. Bennetts, a policy officer with Australia’s peak industry body for urban planning (PIA), underscored AI’s potential to take over low-risk tasks, including development assessments. She acknowledged that AI adds new layers of complexity to the process. She remarked, “There are often more subjective decisions that need to be made, which requires kind of human intervention.”
Regulatory Considerations and Future Implications
As AI’s footprint in urban planning grows, experts like Professor Toby Walsh, Chief Scientist at the University of New South Wales AI Institute, advocate for proactive regulatory measures. Nevertheless, Walsh argues that the Australian federal government needs to step up and lead by setting out the rules that fit with Australia’s distinct regulatory environment. He added that this process would be best informed by developing principles that help to tailor these AI technologies to local needs.
We have to establish guardrails, Ms. Bennetts cautioned. Planning decisions should not be some binary yes or no, but that’s not what history would suggest. She notes that while AI can streamline processes, it cannot replace the critical analysis and community engagement required in more complex scenarios.
Walsh’s speech recognizes the fears about job loss from AI takeover. Yet, he argues that doesn’t need to lead to a reduction in jobs. “If we can reduce that pressure, they can focus more on engaging with communities and creating long-term strategies,” he explained. This view informs our support of Dr. Fu’s vision. He agrees that AI should be used as a tool to make planners far more productive — and not replace planners.
The Path Forward for Urban Planners
With advancements in AI technology already underway in Australia, Dr. Fu and his colleagues aim to navigate the challenges and opportunities presented by these tools. They’re eager to hear how they can help you create the most effective, innovative solutions, aimed directly at Australia’s unique urban planning challenges. Walsh emphasizes this point: “We want to come up with something that works for Australia. The best regulation of the digital space seems to be at the level of the nation-state.”
As urban planners look ahead, there is an acknowledgment that while LLMs can provide valuable support for administrative tasks, significant challenges remain regarding bias and decision-making complexities inherent in planning processes. Ms. Bennetts cautioned against overlooking these complexities: “It’s easy where there’s clear rules, but where it’s a more subjective or trade-off discussion, that’s where it can get more challenging.”
