In a recent tweet, The Drive writer Rob Stumpf asks some important questions about some recent Tesla Full Self Driving (FSD) Beta news. In another tweet, Elon Musk affirmed that the Model 3 and Model Y interior camera was used for driver monitoring for FSD Beta users, and that was a surprise for most.
“We all knew this day would come,” Stumpf said. “However, there are a few things we need to clear the air on.”
Before getting to his concerns, Stumpf detailed a few old tweets by Elon Musk that formed everyone’s prior understanding of what the interior cameras did and didn’t do. In one older tweet, Musk affirmed that the interior camera would be for security purposes for robotaxis and possibly Sentry Mode.
In a 2019 tweet, Elon told us that the interior camera was not activated, and that when it was, people would be notified and given a chance to disable the camera.
I’m not testing the FSD Beta, so some of these questions may have already been answered for those users. If that’s the case, be patient with me and let me know about them in the comments. Please don’t take this as an interrogation for Tesla, angry criticism, or anything like that. I mean no ill will toward anyone with these questions. I just think it would be good for us to know the answer to these.
First off, I think Tesla owners should probably know which users the camera has been activated for. Is it only for FSD Beta users? Is it on for Autopilot or EAP now? Or is it always on for every driver? Knowing this would help everyone to make their privacy decisions accordingly.
Second, I think it’s important for people who it is used for to know exactly how the camera is used. Does a driver-monitoring neural net keep the footage on the local network and decide whether the driver is paying attention, and forward that basic data over the WAN to Tesla? Or does Tesla get the footage and have a chance to review it for accuracy if a driver is tagged as not paying attention?
One side question for the last one: if the NN’s determination is not double checked by staff before someone gets the boot from the program, how sure is Tesla that the determination is accurate? One Black Twitter user is wondering whether facial recognition is inaccurate for people of color as he has seen elsewhere.
A third question I’ve seen multiple people ask is whether the driver monitoring system is an active part of the control loop. In other words, does it cut off the FSD Beta (or Autopilot) if it detects too much inattentiveness? Or is this only something that is later enforced after review?
A fourth question I have is whether FSD Beta participants were notified in some fashion that the interior camera would be monitoring them. If you’re an FSD Beta user, did you know it was being used?
A fifth question: what about FSD Beta, Autopilot, and EAP users who have covered the camera? I know at least one Twitter user above covers it like a laptop camera for privacy concerns. Does a person covering it negate the ability to use any vehicle features?
A sixth question: What data does Tesla retain from vehicle cameras, how long is that data stored for, and who has access to it? Has the company ever been subpoenaed by a government for the data stored (if any)? Can a driver opt out from any vehicle data retention to keep their privacy?
Finally, I’m wondering what security and legal measures are in place to prevent the camera’s misuse by rogue government agencies, hackers, etc. Chances are that they can get past any security measure if there’s a desire to, but it would be nice to know what measures exist.
More broadly, these questions may be appropriate for other vehicle data collected by Tesla’s vehicles. Location data, what exterior cameras record, etc. are all important privacy concerns.
I know many will assert that you have no privacy in your vehicle, as it is traveling on public property in almost all cases. On the other hand, there is still some reasonable expectation below the window level of at least some privacy. If, say, a driver’s crotch started to itch, they wouldn’t expect anyone outside to see that they gave it a quick scratch when nobody was looking, for example.
Also, where one drove and when can be very useful information that could be used against someone in court or in their personal life. Knowing what data was collected, how it is stored, etc. helps people make more informed decisions, even if they’re not doing anything wrong or illegal.
Given that most people thought the camera wasn’t being used, news that it was being used for some users probably feels like a small betrayal for some Tesla owners. It sounds like the camera doesn’t do anything unless one is participating in the FSD Beta program, but at present, we don’t know, and people probably want to know.
This isn’t all about the owners’ interests, either. It may be better for Tesla to answer these questions.
On the upside for Tesla, some drivers may want to cover the camera, which could deprive Tesla of important information for the development of Full Self Driving. If their privacy concerns are addressed, then they may choose to not cover it and Tesla will still get the data they’re wanting to get.
I would imagine that people looking to buy a Tesla would probably want to know more about what the interior camera does, too. Personally, I’d own a Model 3 if it had come out about 3 months earlier. At the time I bought my Nissan LEAF, I was in bad need of a new vehicle for work and the first vehicles hadn’t been delivered. The cheaper versions of the vehicle didn’t come out for a while later, too. I am probably going to buy another EV late this year or sometime next year, and how vehicle data is used is definitely an important question I have.
Featured image: Tesla vehicles charging at a Supercharger station. Image by CleanTechnica.
CleanTechnica's Comment Policy