How AI Can Aid in the Design of Public Safety Facilities

Artificial intelligence (AI) is changing the landscape of design rapidly, offering architects and owners new tools to improve visualization, better communicate with stakeholders and make more-informed decisions. At Little Diversified Architectural Consulting, the internal Emerging Tech CoLab explores technologies that aren’t yet widely adopted (i.e., tools that sit on the edge of innovation or still are experimental). This group plays an instrumental role in testing and developing new tools—focused on AI, automation, digital fabrication and data—that integrate with our internal workflow and help us to work smarter. Ultimately, we’re pursuing efficiencies that allow us to spend more time designing and delivering impactful, meaningful results for clients.

AI & data

AI is everywhere … or is it? You often will see companies add “AI” to their product names when in fact what they’re describing is an automation tool. What’s the difference? Automation is rule-based, such as a formula in a spreadsheet. The computer performs the same task repeatedly, exactly as programmed: no surprises, no learning. AI, on the other hand, learns. It can find connections that we might not even realize exist. At its core, AI excels at analyzing massive amounts of data—far beyond what a human can manage—and synthesizing multiple inputs to discover patterns. It can predict outcomes, optimize based on what it learns, simulate and even generate new content, including text, images and videos.

What AI doesn’t do is empathize, filter, imagine, innovate or make human connections. AI doesn’t design; it serves as an additional voice in the process. Particular to public safety facility design, it’s a tool that enhances and streamlines parts of the design process, deepens understanding of clients, project sites and problems to be solved, and helps to visualize and document ideas faster and more clearly. Architects continue to lead decision-making, to provide direction and to determine what aligns with the project, client and broader vision.

Another buzzword we hear often is data. Everyone is collecting it, most already have it (whether they realize it or not) and most are beginning to understand its value. There are two types of data: structured and unstructured. Structured data is clean and organized, typically found in rows and columns of a spreadsheet. In regard to public safety facility design, it includes demographics, environmental conditions and location-based inputs. AI works well with structured data, because it already is formatted in a way that machines easily understand.

As its name implies, unstructured data is less tidy. In general, it includes emails, photographs, PDFs, social media posts, reviews and any content that doesn’t conform to a consistent format. To make unstructured data more useful for public facility design and design team members, key information first must be extracted and converted it into a structured form.

Both types of data are valuable. Structured data gives us a solid base, while unstructured data fills in the gaps.

Applications for designers

Architecture is a visual profession. What a building will look like must be communicated through drawings and imagery long before construction begins. Hand sketching has evolved into increasingly sophisticated computer-aided drafting and 3D modeling. How does AI enhance visual communication?

With AI, designers can enhance hand sketches, to quickly iterate multiple versions of a concept and rendering, each in a consistent style. This ability allows clients to focus on design differences without being distracted by visual inconsistencies, such as color and shadow. The result is a true apples-to-apples comparison. When used in meetings, this application can help to advance decision-making and bring parties to clear direction faster.

AI also allows for the adjustment of stylization, from loose and sketchy to photorealistic. This flexibility helps to ensure that stakeholders are engaged and can connect with a design on their terms.

Through AI-supported visualization, new images that strengthen and clarify ideas can be generated. For instance, AI can jump start a conversation about how light, color, texture and materials shape a space—and even how that space might feel. Our team uses AI to develop original visuals that are based on blue-sky thinking that are tailored to clients’ specific needs.

AI’s analytical power

By entering a prompt into ChatGPT’s Deep Research feature about a department, in about 10 minutes, a robust, 18–20-page report is generated. It compiles everything that’s needed to be understood when beginning a project, including information about a project’s surrounding community, incident breakdowns by type, community development trends, environmental context and zoning. The reports are sourced with links, which makes it easy to verify statements or flag inaccurate information. The question is, how does a design team use the information? By creating an AI Design Assistant, all the project-related data (i.e., the Deep Research report, program requirements, codes and standards, early design concepts, budget constraints) are compiled. From there, a large language model is fine-tuned. It’s given a name and a purpose, its goals are defined and specific instructions are used to guide it (e.g., the occupants, the building, the design intent). Its boundaries—the do’s and don’ts of how it should respond—also are defined.

Once the AI Design Assistant is ready, particular questions can be asked. For example, what program spaces can be added to support firefighters’ health, well-being and job performance? The responses might include certain strategies, such as sleep optimization pods, a tactical simulation VR lab and a greenhouse kitchen with micro-agriculture. The tool even can be prompted to generate images to help to better visualize these ideas in action.

Further, consider a budget scenario. Perhaps $1 million must be cut. Questions that are posed to the AI Design Assistant could be “What space can be removed with the least program impact, and how can other spaces compensate?” The response that’s provided identifies a candidate for reduction, explains the rationale and evaluates whether that adjustment aligns with overall project goals. If isn’t a fit, the tool offers alternatives or strategies for making smaller cuts across multiple areas to soften the impact. The information that’s generated by the AI Design Assistant is provided in text form, with a supplemental sophisticated table that lists the cost per square foot of the proposed reduction, taking into account how the cost of space varies depending on the use.

The strength of the AI Design Assistant lies in its alignment with goals. It analyzes many variables at once—some that might not have been considered—to paint a more complete picture, helping in making smarter, more holistic decisions.

Deepening understanding

Again, it bears repeating: AI doesn’t design/replace the design process. What it does do is deepen understanding—of the site, the client, the problem. It helps to validate decisions, visualize ideas and work more efficiently. That said, it still is only one part of the team.

At our firm, our work always has gone beyond designing physical spaces. We think deeply about the people who will use them. We consider health, wellness, comfort and effect on the community, to truly go beyond architecture and create spaces that positively effect people and place.