San Francisco is set to become the latest U.S. city to invest in software, created by Texas-based BRS Labs, that monitors and memorizes movements as they are captured on security cameras. The software, AISight, watches footage in real-time and—like a human would—learns to understand, detect, and report “suspicious or abnormal behavior.”
What exactly is defined as suspicious or abnormal behavior? That appears to depend on the environment in which AISight is operating. Its creators say it can be used to flag everything from “unusual loitering” to activity occurring in restricted areas. It could issue an alert after spotting a person leaving a bag unattended in a crowded airport, for instance, or raise alarm if a person is seen trying to cross a perimeter.
San Francisco’s Municipal Transit Authority believes AISight will give it the capacity to track more than 150 “objects and activities” continuously at 12 MTA train stations in San Francisco, according to public procurement documents. BRS Labs has also reportedly struck a deal to monitor the new World Trade Center site in New York. And late last year it was announced that Houston had purchased AISight to be deployed as part of a “citywide surveillance initiative” to “identify potential criminal or terroristic behavioral activity.” It has also been installed in Louisiana for port security, and authorities in El Paso want to use it to monitor water treatment plants near the Mexico border.
The pioneering product has unsurprisingly been lauded by counter-terrorism industry aficionados, but it has caused alarm among privacy and civil liberties advocates. Like surveillance drones, biometric databases, and bomb-proof trash cans, opponents argue, AISight and similar technologies transform citizens into suspects. Because AISight is used to monitor and detect not just acts of crime but potential acts of crime, based purely on a set of algorithms, it is considered part of the push towards pre-emptive—or “pre-crime”—policing, which treats everyone as a potential criminal and targets people for crimes they have not yet committed (and may never commit).
For years researchers have been trying to develop advanced “intelligent” surveillance technology of this kind. The European Union previously ploughed €2.5 million ($3.1 million) into a project called Samurai to “develop a real-time adaptive behaviour profiling and abnormality detection system.” And the Department of Homeland Security has even been building a program called Future Attribute Screening Technology that it hopes will "detect cues indicative of mal-intent" based on factors including ethnicity, gender, breathing, and heart rate.
Now that the technology is beginning to hit the marketplace, there is likely to be a sales boom. San Francisco alone plans to spend $2 million on AISight. A report last year by the Homeland Security Research Corp., predicted that this decade will see a fusion of CCTV with biometrics and “behavioral suspect detection”—a market it estimates will experience growth from $750 million in 2011 to a massive $3.2 billion by 2016. So when BRS Labs boldly boasts that AISight is “a revolutionary product that has changed the security industry forever,” it’s hard to disagree.