Technology upgrades to improve the pedestrian experience

LLMs and AI are the latest technologies that should be embraced by urbanists.

Every state, regional, or local department of transportation claims to accommodate walking. After all, it’s the fundamental mode of transport. But anyone who tries walking in modern America knows what a gamble it can be. The National Highway Traffic Safety Administration has reported a 13% increase in pedestrian deaths and a 2% increase in cyclist fatalities in 2022. 

One of the key factors exacerbating the safety concerns for pedestrians and cyclists is the lack of proper infrastructure. It’s hard work to keep up with all the street deficiencies, let alone prioritize which projects should be funded first.

Avoid the lazy critiques of technology

I’ve written extensively about how to design streets to calm down drivers. I’m a big believer in using technology to improve the human experience. It puzzles me that so many journalists and urbanists have opted for a “tech can’t save us” approach to life, when there’s a ton of “tech can absolutely save us” evidence. 

Language, for example, is a form of technology that is capable of harm or benefit. It’s a tool. A large language model (LLM) is a computational powerhouse of language. A couple of Vision Zero advocates talking about ideas to reduce crashes aren’t just squawking random sounds out of their faces—they’re using the technology of language.

Every week you’ll read more hype and doomsday headlines about the power and threats of LLMs and artificial intelligence advancements. Before I get into positive ways to use LLMs and AI, here are some definitions so we’re on the same page.

AI basics:

AI is a field of computer science that focuses on creating intelligent hardware and software capable of performing tasks that typically require human intelligence. AI is what makes it possible for computers to learn from data, make decisions, and perform tasks without explicit programming for each specific task. Lightning-fast pattern recognition is a major part of this.

LLM basics:

LLMs are computer programs designed to understand and generate human-like text. They are trained using massive amounts of data, such as books, articles, and websites, to learn the patterns and structures of language. Just like how we learn from reading and listening, LLMs learn from data to become really good at understanding and producing written text.

One downside is that they pull answers from so much text, that the answers are often wildly inaccurate. For example, if an LLM’s data set included 5 reports on why eating meat is good and 5 reports on why eating meat is bad, the LLM will blend the information and return a nothingburger. Still, LLMs are incredibly powerful. And contrary to what you might’ve read, they’re incapable of emotions. (I didn’t say that makes them traffic engineers, you said that.)

When you read or hear the latest “technology is the enemy” narratives, I want you to remember:

7 ways city planners can leverage LLMs and AI to make walking safer and more appealing

1. Predict pedestrian flow and demand 

LLMs can predict walking patterns and identify areas with high demand by analyzing a combination of historical data, weather conditions, special events, and other likely influences. This would empower city planners with the best possible “desire lines” analysis to prioritize sidewalk projects, crosswalks, speed tables, bump-outs, etc.

2. Assess walkability and safety 

LLMs can analyze diverse data sources to evaluate the walkability and safety of pedestrian infrastructure. Think crash reports, street design features, and community feedback surveys. This analysis helps identify areas that need better lighting, traffic calming measures, ADA ramps, electric scooter parking hubs, etc. LLMs can process tons more information about pedestrian behavior, desire lines, and vehicle conflict points. Engineers can use LLMs to adjust signal timings for pedestrian priority to reduce wait times, especially at busy crossings or corridors.

3. Personalize pedestrian route planning 

Real-time data on weather conditions, pedestrian traffic, and user preferences can be analyzed by LLMs to generate personalized walking routes. Obviously, this gets even more important as local agencies improve connectivity between mixed use developments and you have interesting places to walk. LLMs will be able to consider individual factors and provide tailored recommendations, hopefully encouraging more people to choose walking as a practical and enjoyable mode of transportation for short trips.

4. Create more accessible pedestrian infrastructure

LLMs can analyze data related to accessibility barriers, such as steps, curbs, and narrow sidewalks, to design more inclusive and accessible pedestrian infrastructure. This enables city planners to identify areas that need modification and make adjustments to ensure easier mobility for individuals with disabilities and other mobility challenges.

5. Encourage more people to walk 

LLMs can assist in developing targeted marketing campaigns and initiatives aimed at encouraging more people to embrace walking. Transit agencies would benefit from this type of application, clarifying first/last-mile paths to bus stops. Transit planners can raise awareness and provide incentives to inspire a mode shift from drive alone to a walk/bus combo.

6. Make cities more sustainable 

LLMs can identify and support the implementation of sustainable strategies, such as reducing car dependency and increasing green spaces. This could also be useful for nonprofit and advocacy organizations trying to synthesize gobs of data to improve quality of life in the built and natural environment.

7. Create more livable cities

Cities and towns are more enjoyable places to live, work, and spend money when they prioritize walkability, bikeability, and accessibility. LLM-driven improvements contribute to an enhanced quality of life, vibrant communities, and an increased sense of connection between people and their urban environments.

Technology is a big reason that things get better in the end.

But wait, there’s more…

Here are some examples of organizations putting this technology to work in city planning:

New York City Department of Transportation is using LLMs to to analyze traffic data and identify patterns that can be used to make better decisions about traffic management.

California Department of Transportation is using LLMs to develop a new traffic management system that will collect and analyze traffic data in real time, and then use that data to make decisions about traffic signals, lane closures, and other traffic management measures.

The World Bank is using LLMs by analyzing traffic volumes, demographics, and other factors to help cities make decisions about how and where to invest in their transportation infrastructure.

Uber & Lyft are using LLMs to predict demand for rides, optimize routes, and improve customer service.

Here are some practical ways these organizations are using LLMs:

Summarizing large amounts of text from news articles, research papers, and legal documents.

Rewriting text in different styles for marketing copy, translating languages, and creating personalized content.

Extracting information from text to find gold nuggets, identifying key points of large reports, and creating a taxonomy of topics.

Searching for text that is similar to a given query within mountains of reports and transcriptions.

Clustering text into groups based on similarity to find trends and suggest recommendations.

Classify text into categories, like tagging documents, identifying spam, and classifying customer sentiment.