Tesla workers shared sensitive images recorded by customer cars

 Tesla Inc. tells the countless owners of electric vehicles that "privacy is and will always be enormously important to us" in LONDON. It states on its website that the cameras it incorporates into automobiles to aid driving are created from the ground up to secure your privacy.

According to interviews conducted by Reuters with nine former employees between 2019 and 2022, however, groups of Tesla employees privately discussed via an internal messaging system occasionally extremely intrusive films and photographs taken by customers' car cameras.

A few of the recordings showed Tesla customers in awkward circumstances. A video of a man approaching a vehicle while entirely naked was described by one ex-employee.

Accidents and cases of road rage are also shared. According to a different ex-employee, a Tesla was seen in a crash video from 2021 traveling quickly through a residential neighborhood before striking a child riding a bike. Both the toddler and the bike flew in opposite directions. According to the former employee, the video spread like wildfire through private one-on-one talks within the Tesla office in San Mateo, California.

Other pictures were more commonplace; for example, staff added witty comments or commentary to images of dogs and funny road signs to turn them into memes before uploading them in personal group conversations. According to multiple former employees, some postings were only visible to two individuals, while others were accessible to hundreds of them.

Tesla claims that its camera recordings are anonymous and are not connected to "you or your vehicle" in its online Customer Privacy Notice. However, seven ex-employees told Reuters that a computer system they used at work would be able to expose the location of recordings that could perhaps reveal the residence of a Tesla owner.

Additionally, according to one ex-employee, several recordings appeared to have been captured while automobiles were parked and switched off. Tesla used to be able to obtain video records from its vehicles even when they were not in use if the owners gave permission. Since then, it has ceased to do so.

Another ex-employee claimed, "We could see inside people's garages and their private properties." Let's imagine a Tesla customer had something unique in their garage. People would share those kinds of things, right?

Detailed inquiries submitted to Tesla for this article went unanswered.

According to two persons who watched it, a video of an unusual submersible vehicle parked inside a garage was discovered and shared by several employees about three years ago. The white Lotus Esprit sub, known as Wet Nellie, appeared in the 1977 James Bond movie The Spy Who Loved Me.

Elon Musk, the founder and CEO of Tesla, purchased it at auction in 2013 for approximately $968,000. Whether Musk was aware of the film or that it had been circulated is unknown. A comment from Musk was not forthcoming.

Reuters contacted more than 300 former Tesla workers who had worked there over the previous nine years and were engaged in the creation of its self-driving system in order to report this story. Over a dozen individuals volunteered to answer questions under the condition of anonymity.

Ex-employees claimed they hadn't saved the shared films and photographs, but Reuters was unable to obtain any of them. The news organization was also unable to determine whether or to what extent the practice of exchanging recordings, which was observed in some areas of Tesla as late as last year, is still practiced today. When contacted, some former workers claimed the only sharing they saw was done for proper business needs, including asking coworkers or managers for help.

Sharing private recordings serves to highlight one of the less-noticed characteristics of AI systems: They frequently need armies of people to assist in teaching machines to master automated operations like driving.

In order to train its cars to recognize pedestrians, street signs, construction vehicles, garage doors, and other items encountered on the road or at customers' homes, Tesla has hired hundreds of individuals in Africa and later the United States to categorise photographs. To do that, automobile cameras were used to record thousands of movies or photos that data labelers could see and identify objects in.