Using Generative AI to Draft Zoning Codes
Zoning Practice — October 2023
By Norman Wright, AICP
Not a member but want to buy a copy? You'll need to create a free My APA account to purchase.
Create account
Artificial intelligence (AI) has been a topic of conversation for many years now. Yet, it was the advent of new consumer products like ChatGPT that helped the broader public see how disruptive (and transformative) the technology can be. Five minutes with any of the large language models available online will quickly demonstrate the incredible speed, accuracy, and comprehensiveness of information these tools produce. None of the models are perfect, of course, but limitations are often greater than we realize.
Yet, it is clear that artificial intelligence tools — specifically, the kind that produces generative text and images — are here to stay. The most popular products receive enormous amounts of investment and are on a rapid trajectory for mainstream usage in every industry. Planning is no exception.
This issue of Zoning Practice introduces key concepts related to the common use of today's generative AI models and then provides an overview of experiments done to test the current strengths and weaknesses for developing zoning policies at the local level. As you'll see, there is a lot for us all to learn — including the models themselves. The current products have limited abilities, but the potential is boundless.
Details
About the Author
Norman Wright, AICP
Norman Wright, AICP, is the founder of Parameter, a design and analytics consultancy that specializes in creating plans, policies, and best practices for local government. Prior to this, he was a local government executive leading planning, economic development, and community development efforts in Oregon, Colorado, Tennessee, and South Carolina. His teams have delivered award-winning work recognized by the American Planning Association and Urban Land Institute. He holds a Master’s degree in City and Regional Planning from Clemson University.