City of San Jose testing use of AI to identify potholes, graffiti and homeless encampments
SAN JOSE, Calif. - In Silicon Valley, tech companies are already harnessing the power of AI. Now the city of San Jose wants to do the same.
"It's a competition. Let them demonstrate what they can do," says Khaled Tawfik, the Chief Information Officer for the City of San Jose.
What they want AI to do, is to identify small problems, before they become big ones. They're looking at everything from potholes and graffiti, to homeless encampments.
Blue Dome Technologies is one of five companies participating in the pilot program. They showed us their technology at work. Right now, they are primarily focused on road hazards, but they're training their model to identify other things like growing homeless encampments.
"What we use is a very simple Go Pro camera mounted on city vehicles and when they drive around they'll upload the video. And over time we find out how the encampments have grown," says Swaroop Patnaik of Blue Dome.
Other companies like Ash Sensors are working on this too. They use a combination of cameras, lasers and sensors. And while their focus is primarily on the structural stability of buildings, they are also trying to meet the city's needs.
"Because that's one of the things with AI, you have to teach it what to go look for. So that's a pothole. This is a building. Here's graffiti, here's not graffiti," says Kenneth Salsman, Ash Sensor's Chief Technology officer.
The city of San Jose has just one car outfitted with cameras, and it's only done three test runs so far. But homeless advocates have concerns, that this high-tech solution won't solve a very human problem.
"The public's saying, 'Oh it's another tool, but a tool for what?' You've got to get unhoused people into housing. That's the solution. It's always been the solution," says Scott Wagers, a homeless advocate.
And city officials say they are very conscious of potential pitfalls and that's why policy and transparency around AI is crucial.
"There is a privacy risk. There is bias risk that comes with AI. And this is really a key issue we're trying to address because if we jump to conclusions without learning. If we tried running without walking, we get ourselves in trouble," says Tawfik.
There is no set timeline for how long this pilot will last. City officials say it will take a while to see what AI can do. At the end, there will be a competitive bid process, assuming they decide to proceed.
They also say the information that comes out of this pilot program could be a model for other cities in the future.