Can artificial intelligence offer a solution to make zoning more rational and reasonable?
At one time or another, most architects and developers have found themselves frustrated by zoning regulations. Zoning is not an element of design, but a tool used to ensure that our cities are built in predictable ways so they can provide adequate services and infrastructure for their occupants, as well as protect the interests of neighboring property owners. However, as our cities and society have grown more complex (and litigious), zoning has likewise grown to be more complicated.
As a result, zoning can place limiting parameters around creativity in design and development. For example, a meticulously crafted site plan can be met with contention from a plan examiner over an interpretation of zoning text, leading to either a significant redesign of the project or a costly legal appeals process. Or, a project’s viability can hinge on minor variances from zoning requirements, such as a setback dimension differing from zoning mandates by a few feet. In such instances, can artificial intelligence offer a solution to make zoning more rational and reasonable?
As AI has gained popularity, many groups have developed platforms that developers and architects can use to automate zoning analysis and iterate design solutions. With these platforms, one can enter a street address and instantly receive multiple design and development options for the specific site. These tools are impressive and useful, but they currently only capture today’s zoning rules, and therefore are not fully harnessing AI’s potential to improve zoning and advance larger city planning goals. In the next generation of AI zoning tools, shifting the target market from architects and developers to cities and municipalities would allow zoning to shift dynamically in real time.
With AI tools, cities could evaluate proposed buildings based on how they meet specific performance standards, expediting the variance approval process and creating a more adaptable system. Take, for example, New York City’s 15-foot setback regulation. This zoning ordinance creates a universal, city-wide rule that buildings of a certain height must be set back 15 feet from the street to preserve daylight access. Frequently, buildings function better when designed with smaller setbacks. AI tools could assess and approve or deny variance requests based on performance-based criteria, such as maintaining adequate daylight penetration to the street. Existing tools, like the Daylight Evaluation in New York City’s Midtown Manhattan zoning resolution, could provide a foundational framework for these new AI tools.
As codes in cities and municipalities nationwide expand and begin to overlap, we have created a complicated regulatory environment that produces inconsistent outcomes. Despite efforts from building departments across the country that strive to interpret and enforce regulations consistently, there nevertheless remains a small, but significant level of discretion. By creating an AI powered plan examiner to evaluate code compliance for buildings, we can ensure that all rules are applied equally. This would also allow designers to iterate and check their work, expediting the design process and establishing a more predictable permitting process. Plan examiners could then spend less time reviewing drawings and more time refining the AI model or confirming that projects are built as designed.
While most cities and municipalities have existing processes through which zoning can be modified to suit a project’s needs, they can be prohibitively expensive and time consuming, requiring a full suite of consultants, and therefore are typically only considered for larger projects. AI tools for minor zoning modifications could make discretionary actions more accessible to smaller projects. For larger projects, they could support the evaluation of more substantial variances, such as increased floor area or change of permitted uses.
While it is important for policymakers and elected officials to have approval authority for these projects, AI can streamline the environmental review process required for zoning modifications, leading to more informed decisions. These tools can also be used in stakeholder outreach and participatory design, allowing constituents to adjust and visualize the impacts of various zoning criteria at reviews.
While current AI tools offer valuable assistance in navigating existing zoning regulations, the promise of AI in zoning and city planning lies in its ability to dynamically adapt to evolving urban needs. By shifting the target users of zoning AI tools from architects and developers to cities and municipalities, AI can empower decision-makers to create more flexible and responsive regulatory frameworks, streamline the permitting process, and ensure equitable code enforcement. Ultimately, these tools can allow buildings to meet the needs of the cities they are built within.
Ben Abelman, AICP, LEED GA is senior associate and director of zoning and predevelopment at FXCollaborative. Connect with him on LinkedIn.
