- a cybersecurity expert has created "surveillance detection" technology he's calling surveillance detection scout, optimized for use with tesla vehicles.
- the tech works through the camera feed in tesla's sentry mode security system when the car is parked, but it is not affiliated in any way with the automaker itself and could be adapted to other vehicles using dash cams.
- what's the point? to know if you're being followed or if someone's trying to steal your car.
in some places, a tesla is still a rare enough sight that people will stop to take pictures of it. turns out, there's one tesla model s that might be taking pictures right back.
the model s in question is one that has spent some time with truman kain, a senior information security analyst at the cybersecurity firm tevora. kain presented a surveillance and countersurveillance framework called surveillance detection scout at def con 27 in las vegas last weekend. surveillance detection scout is a collection of available software (including kain's own work) and some additional pieces of hardware—a jetson xavier minicomputer from nvidia and a usb drive—that taps into three camera feeds in any tesla electric vehicle and processes the visual data in real time to identify if someone is trying to steal or break into the vehicle.
scout, as kain calls it, does this by using facial recognition software and an automatic license-plate reader, both of which are outlawed in some states. if scout detects a face or license plate repeatedly, it notifies the driver in the car and through a phone alert.
kain has been plenty public about his work. he says that one of the main reasons he developed scout was to draw attention to the kinds of surveillance that are possible with technology available today. he also thinks it could be used as a tool by individuals who are running surveillance detection routes to see if they're being followed. speaking to car and driver, kain said he thinks raising awareness is the best way to keep people secure.
"imagine two scenarios. one, we're where we are at now with me having just released this and someone next month basically repackages what i created, offers free or low-cost cloud storage to users, marketing it like 'nest, for your tesla,' " he said. "it's likely that people would start blowing the whistle right away . . . because they are now, if they weren't already, educated regarding the implications of driver data being collected en masse by a third party.
"[in] the second scenario, i never released this tool publicly or spoke about it and the same thing mentioned above happens. i believe that thousands of drivers would opt in without a second thought, just like most of us have [done] to facebook and other 'free' services we use every day that in reality are living off our data."
even though kain used a model s for scout, there's nothing preventing someone from doing something similar with dash cams in another car, which kain acknowledges. the differences would be, first, that scout uses the cameras that come with tesla's sentry mode (similar to a home security system) to monitor the car when it's parked and off and, second, using simple dash cameras to save clips to a flash drive would mean that you wouldn't be able to analyze the data until you send it to a computer later.
"with scout, the inference and detection starts as soon as the file is written to the drive," kain said.
kain didn't just present the scout to the world at def con. he also uploaded the software to github, and he thinks that your average tech-savvy person could get their own scout up and running in roughly three hours for between $300 and $1000, depending on how fast they want the detection to take place and the size of the onboard storage.