Police in the US state of Delaware in the east of the country are preparing to deploy “smart” cameras in their vehicles to help authorities detect a vehicle carrying a fugitive, a missing child or a person disoriented aged.
The videos will be analyzed by artificial intelligence software that can identify vehicles by license plates or other features, and thus give “extra eyes” to patrolling officers, says David Hinojosa of Coban Technologies, the company that provides this material.
“We are helping police officers stay focused on their work,” he says, calling the new technology “steroids camera”.
The program is part of a growing use of the alliance between artificial intelligence and video to fight crime, a trend that worries privacy advocates and civil liberties.
The US start-up Deep Science uses the same technology to help traders detect real-time armed robbery, identifying weapons or masked attackers, triggering automatic alarms.
According to Sean Huver, the co-founder of Deep Science and a former Pentagon engineer, this technology is more efficient and less expensive than human security guards.
“The problem with the security guards is that they are bored,” he says.
– Schools, hotels, shops –
Computer vision technologies are already being used for autonomous vehicles or drones to recognize and interpret the environment, says Saurabh Jain, product manager for Nvidia’s computer graphics group, which produces computer chips for these new systems of security.
Nvidia already has around fifty partners using its technology. One of them, Umbo Computer Vision, based in California, has developed a surveillance system that can be used in schools, hotels or other places, analyzing video to detect intrusions and threats in time real and send an alert to the security guards.
The Israeli startup Briefcam uses similar technology to interpret CCTV footage.
“A video is not structured, you cannot search it,” says Amit Gavish, CEO of Briefcam in the United States. Without artificial intelligence, he says, “you’d have to watch hundreds of hours of footage using only fast forward and fast back.”
“We detect, track, extract, and classify each object in the video to make it a database.”
This may allow investigators to quickly find suspects in CCTV images, a system already used by law enforcement agencies in hundreds of cities around the world, including Paris, Boston and Chicago, says Gavish.
“It’s not just time-saving, in many cases it would not be possible because people who watch video images become ineffective after 10 to 20 minutes,” he says.
The Russian start-up Vision Labs uses Nvidia’s facial recognition technology to identify shoplifters or troublemakers in casinos.
“We can deploy (this technology) anywhere,” enthuses Vadim Kilimnichenko, Project Manager at Vision Labs.
His clients also include banks for which facial recognition can help determine if someone is using a false identity.
– Risks for privacy –
For Marc Rotenberg, president of the Electronic Privacy Information Center, the rapid growth of these technologies raises privacy risks and calls for regulatory control over how data is stored and used.
“Some of these techniques can be helpful, but there are huge privacy issues when systems are designed to capture identity and make a decision based on personal data,” worries Marc Rotenberg.
“This is where the problems of profiling, bias and precision come into play.”
The use of artificial intelligence systems in court must be subject to legal guarantees and transparency, he says.
In a blog post earlier this year, Shelly Kramer of Futurum Research said artificial intelligence holds great promise for police forces, whether it’s for surveillance, identifying threats on social networks or use of “bots” to detect lies.
“This encouraging promise, however, comes with a host of risks and responsibilities,” she warned.