Privacy, Security, and Trust Emerge as Key Themes at 2024 Innovation Summit

Experts from government, academia, industry, and the local community gathered in Morningside Heights on April 16 to continue discussions on advancing community-driven, hyperlocal smart city technology. In keynote lectures, panel discussions, lightning talks, and pitch sessions, the attendees discussed a diverse array of technical innovations and policy considerations necessary for implementing smart streetscapes.

Andrew Smyth, director of the Center for Smart Streetscapes (CS3) and the Robert A.W. and Christine S. Carleton Professor of Civil Engineering and Engineering Mechanics at Columbia Engineering, emphasized the importance of collaboration for CS3.

“The CS3 Innovation Summit united researchers, leaders, public officials and students to advance our vision of innovation built through partnerships to improve life on the streetscape for all,” he said. “With ‘partnerships’ being a key element, this summit is a huge milestone in strengthening and growing the partnerships that make our vision a reality.” 

Dmytro Pokhylko, Innovation Ecosystem Director at CS3, pointed to one type of partnership the center helps enable: student-led initiatives. In introducing student research presentations, Pokhylko emphasized how CS3’s approach impacts student research. 

 “The accelerator’s robust programming — coupled with community and industry engagement — supports students through the process of validating their innovations and positioning them toward greater societal good,” he said. 

Pursuing safety and security

The use of technology to ensure safety — particularly for pedestrians — was a key theme at the summit.

In his keynote presentation on harnessing AI in smart streetscape technology, Jumbi Edulbehram, NVIDIA’s head of global business development for Smart Cities and Spaces, offered several examples of instances where real-time data collection and edge computing are already being deployed to make streetscapes safer. 

He pointed to the startup NoTraffic, which instruments roadways with sensors to deploy its autonomous traffic management system. Edulbehram shared data showing marked decreases in traffic congestion in Tucson, Ariz. At 84 intersections, accidents related to red light running have decreased by 41 percent since the technology was deployed.

“Once you collect information on these kinds of scenarios and understand how they happen, then you can instrument the real world to make sure these tragedies don’t happen,” he said, referring to accidents between a vehicle and a pedestrian, cyclist, or someone riding a scooter.

Several of the day’s speakers pointed out that these systems aren’t without their dangers. 

In the panel discussion, “Privacy & Security in Smart Streetscape Technology,” Columbia Engineering Computer Science Professor Salvatore J. Stolfo observed that new hardware increases the attack surface, making it possible for adversaries not only to siphon off data but also to manipulate the systems designed to increase safety.

“If I can get access to a device, I can make that device misuse its sensors,” said Stolfo, who is a pioneer in internet privacy and security. “I can tell it that there are a whole bunch of citizens walking across the street when there are none, or vice versa. What if I told all of the traffic lights to turn green at the same time?”

The solution, he said, is developing and deploying these systems with an overarching security architecture.

“The questions I ask my colleagues working on streetscapes are, Who is going to manage security? Who's going to be responsible for ensuring the systems are going to be configured correctly? Are they going to be tested?”

Communication is critical

While the systems advanced by the Center for Smart Streetscapes promise increased safety and convenience, previous iterations of this type of technology have sparked controversy. Two of the center’s research thrusts — Public Interest Technology and Security, Privacy, and Fairness — are focused on developing technologies with privacy by design and community engagement standards that strengthen privacy and community support. 

Rachel Cummings, associate professor of industrial engineering and operations research at Columbia, described the crux of this set of issues.

“The easiest way to have a smart streetscape is by putting out a bunch of devices that collect data on the public,” she said. “Incorporating privacy is a big challenge.”

Kenji Yoshihira, Director of US Ecosystem Innovation at NEC Laboratories America, Inc., described a successful case in which a consortium, including his company, implemented smart streetlights in the public streets of commercial districts in Roppongi, Tokyo. These smart streetlights utilized facial recognition technology, but the purpose was solely to observe the movement of anonymous individuals, with no personal data being recorded, ensuring complete privacy protection.

This initiative emphasized the importance of addressing underlying civic sentiments and meticulously explaining the value and benefits of the technology to all stakeholders beforehand. The project involved early engagement with local government, law enforcement, academics, commercial councils, and unions.

“This was a really successful case, and we saw tremendous improvements because of this approach to communication with the local community,” he said.

Stolfo summed up many of the participants’ sentiments about trust and community acceptance.

“Convincing the public is the hard part,” he said. “The easy part is breaking the trust.” 

Throughout the day, panelists and speakers continually made the point that building that trust rests on bringing community stakeholders on board early in a project and taking their concerns seriously. 

Cummings summed up the sentiment during her panel.

“Many folks in this room are very thoughtful about making sure that technologies on the street provide privacy and security guarantees, but if people don’t know about those guarantees then they won’t feel good about the technology.” 

About the Center

The mission of the Center for Smart Streetscapes (CS3), a Gen-4 NSF Engineering Research Center founded in 2022, is to forge livable, safe, and inclusive communities through real-time, hyperlocal streetscape applications built on advancements in edge-cloud technology, wireless-optical engineering, visual analytics, computer security, and social science.

The center strives to co-develop, with the involvement of the stakeholders at all levels of the design process, a platform for technological solutions that can address current and future community-identified problems, such as vulnerable users' safety, traffic efficiency, public safety, and hyper-local environment sensing.

Researchers are working closely with a diverse cohort of more than 80 non-academic community stakeholders – industry partners, community organizations, municipalities, and K-12 schools – both as collaborative co-producers of knowledge and as auditors of technology research and development.