Friday 17 April 2026
An enthusiastic crowd of researchers, industry leaders, public-sector innovators, and community stakeholders gathered at the Forum on Columbia University’s Manhattanville campus recently to discuss the difference between AI hype and the tangible value that technology can bring to the public.
Hosted by the Center for Smart Streetscapes (CS3), an NSF Engineering Research Center, the event featured keynote addresses from Kristen White of Google Public Sector and Chris White of NEC Labs America, as well as panel discussions, student pitch competitions, and a poster session.
In his opening remarks, centre director Andrew Smyth explained how CS3, now in its fourth year, has shifted its focus to stay ahead of technological advancement.
“The AI boom has radically changed the research landscape, leading us to pivot our research agenda,” said Smyth, who is Robert A W and Christine S Carleton Professor of Civil Engineering and Engineering Mechanics.
Today, CS3 focuses on AI-driven reconfigurable surfaces, on-sensor AI, multi-modal sensing, agentic AI, AI governance, and the incorporation of community feedback in decision-making about streetscape technology and policy.
Smyth emphasised that the centre’s research, partnerships, and community involvement activities support its overarching mission to forge liveable and safe communities through real-time, hyper-local streetscape applications.
While many research centres focus solely on technical metrics, Smyth noted that CS3 is “probably unusual as a technological centre in that we have a statement of values”. Central to these values is the conviction that “tech should serve the public good” and that engineers must be trained to “make their work publicly legible”, he said.
From sensors to systems: Real-time traffic analysis for faster decision-making
In the morning panel discussion, CS3 managing director Olivia Moore moderated a conversation about a collaboration using new data streams and analysis techniques to turn raw traffic data into actionable traffic insights. The panel featured Robert Holbrook of the NYC mayor’s office, Ahmed Darrat of data company INRIX, and Mehmet Turkcan of CS3.
In a preliminary talk at the beginning of the session, Turkcan demonstrated how CS3 researchers developed a system that uses physics-constrained graph neural networks to make sense of the city’s traffic problems by incorporating video data from street-level cameras and GPS data from individual vehicles for calibration and validation. This hybrid approach seeks to solve the 'garbage in, garbage out' problem of modern urban modelling, where poor input data make it difficult for traffic planners to trust outputs from today’s models.
“By calibrating our model of traffic flow across New York using this video data, we can get accurate results in regards to the actual traffic flow in the city,” he said.
In the discussion, Ahmed Darrat of INRIX noted that integrating its own GPS data with video footage significantly increased the utility of both data streams.
“A hybrid of the two is actually probably the best solution because you get the best of both worlds, tracking the subtleties of traffic flow everywhere all the time,” he said.
Robert Holbrook emphasised that actionable traffic data is "foundational to everyone’s life in New York" because it enables the city to manage traffic in real time. For example, the city employs the Midtown in Motion traffic management system to adjust traffic signal timing in the Central Business District. The city also uses traffic data to make decisions about curbside operations that affect pedestrians all over the city, such as 'daylighting' corners by replacing on-street parking with curb bump-outs, allowing safer pedestrian crossings.
Testbed perspectives on local value of technology in streetscapes
Panelists, l-r: Olivia Moore of CS3, Mehmet Turkcan of CS3, and Robert Holbrook of the NYC mayor’s office.
The afternoon sessions shifted focus towards how these innovations are applied across various environments, ranging from the streets of West Palm Beach to the highly controlled 'mini-city' that is an international airport. Moderator Bill Kenworthey of HOK led a discussion with Jason Hallstrom of Florida Atlantic University, VJ Rajvanshi of Middlesex County, and Brian Cobb of Cincinnati/Kentucky International (CVG) airport on bridging the gap between high-tech research and ground-level public value.
According to Hallstrom, who works closely with officials from the City of West Palm Beach to operationalise their smart streetscapes technology testbed, the city measures the value of smart technology by its impact on community liveability and trust. He emphasised that technology must serve the people first, rather than being an end in itself. “It always begins with community and making sure that the technology is actually meeting a need,” he noted.
In Middlesex County, the focus is on the scale and longevity of infrastructure. VJ Rajvanshi discussed the 'Data City' initiative in New Brunswick, which utilises a 3.5km corridor of connected intersections to improve safety. Rajvanshi highlighted the shift from reactive to proactive governance through data.
“We are trying to move from a place where we are looking at things after the fact to where we can identify traffic patterns before they escalate into problems and focus our attention to where it matters most,” he said.
Brian Cobb offered a unique perspective from CVG airport, which serves as a specialised testbed where technology can be validated before hitting city streets. Cobb described how autonomous floor scrubbers at the airport do more than just clean: they act as mobile sensor hubs that continuously monitor conditions at the airport, including security concerns such as unattended baggage.
“It’s not just a cleaning bot, it’s also a security bot because you're running cameras all the time and a mapping utility because it’s constantly surveying the airport,” and potentially providing real-time data streams to other applications.
Looking ahead
The day demonstrated that effective deployment of AI requires close partnerships combined with technological know-how. In his closing remarks, Andrew Smyth discussed how the day’s discussions successfully bridged the gap between theoretical research and the practical needs of infrastructure owners, operators, and planners. By bringing together global industry leaders like NEC Labs and Google with local stakeholders like West Palm Beach to New Brunswick, the event reinforced CS3’s role as a vital hub for public-private innovation.
As the centre moves into its fourth year, CS3 remains focused on ensuring that technological leaps in AI translate into tangible public value. As Smyth noted in his final remarks, the centre's ultimate success lies in its ability to listen to community needs before engineering begins, ensuring that the "purposeful, vibrant streetscapes" of the future are built on a foundation of trust, safety, and public good. By maintaining this commitment to a core statement of values, CS3 aims to move beyond simple data collection to create urban environments that serve the people who live in them.