Podcast: Play in new window | Download
Episode 29: Tesla Data Storage – Show Notes
Tesla has given the world a new promise between its electric engine, sleek styling, and amazing autopilot. But under the hood lurks a dangerous secret. Every time a phone pairs with the vehicle, all of the contact data, phone call logs, messages, and more gets pulled into the Tesla’s memory. All it takes is a cable and a little data knowledge to get access to everything – including a trove of unencrypted personal information. Today we discuss the dark side of convenience with Tesla data storage in its cars.
Additional Links on Tesla Data Storage
Unencrypted Video and Personal Data Stored on Teslas Raise Significant Privacy Concerns Report from The Drive detailing what was found on wrecked Model 3 vehicles and some background on other Tesla privacy concerns.
Researchers Find Mountains of Sensitive Data on Totalled Teslas in Junkyard Additional article from BoingBoing on Tesla data storage findings from those seeking the bug bounty on research.
Tesla Data Storage Episode TranscriptView full episode transcript
Welcome to the Data Science Ethics Podcast. My name is Lexy and I’m your host. This podcast is free and independent thanks to member contributions. You can help by signing up to support us at datascienceethics.com. For just $5 per month, you’ll get access to the members only podcast, Data Science Ethics in Pop Culture. At the $10 per month level, you will also be able to attend live chats and debates with Marie and I. Plus you’ll be helping us to deliver more and better content. Now on with the show.
Marie: Hello everybody and welcome to the Data Science Ethics Podcast. This is Marie Weber
Lexy: and Lexy Kassan
Marie: And today we are going to be doing a quick take on some of the news around the data collection policies that Tesla currently has on its vehicles. So to kind of queue up this story, we have an article that we’re going to be linking to from the drive. It talks about a few different people who have been purchasing wrecked Tesla model threes and they’re basically taking these cars and they’re looking at the data that’s been stored on them. Part of the reason why they’re looking at this is because Tesla has a bug bounty program, which means that if people can find bugs that Tesla has in their system, Tesla will pay them money for basically doing that white hat hacking and then they can use that information to make their system better. But it also means that the people going in and looking at these cars were able to uncover some very surprising things in terms of the amount of data that Tesla has been storing on their vehicles and the way they’ve been storing it, Aka it’s not encrypted.
Lexy: Yeah. And there’s a bunch of personally identifiable information on there, not only of the people who were in the vehicle, but of everyone they know.
Marie: It’s not just recording what’s happening as the vehicles driving and who’s driving type of thing, because it can connect to digital devices. It can then look at information on those devices. So the scope of what was being accessed by these vehicles from the personal devices included contacts from foam books, calendar entries, and even email addresses.
Lexy: Right. So when you think about the triangulation of personally identifiable information, you’ve got all of it. You could have where someone was, when, what their phone number is, their name, their email address, and you have a personally identifiable link that you could leverage. That’s a very sensitive piece of information.
Marie: The cars were also recording where people were going and they were keeping this information for quite some time. The vehicle that they looked at said that the researchers were able to discover the last 73 locations that the driver had navigated. Yeah,
Lexy: yeah. The other thing that was very interesting about this specific vehicle is that it was a company vehicle and they noted that the devices that paired with the vehicle, they had 11 different devices that paired with the vehicle. So therefore potentially 11 different drivers of that vehicle. So if you think about this as a fleet vehicle, you now have the personal information of 11 people, their contacts, their calendars, all of them, all of the information of all the places. They went off everything and it was unencrypted on the machines. Within the vehicle?
Marie: No. As we mentioned, these were salvage vehicles, so they were vehicles that had been crashed and the other thing that was recorded was actual videos of the crash happening. So it is understandable in some regards why Tesla has incorporated some of these features into their vehicle. They want it to be a great experience. They want people to be able to pair their devices and to be able to take calls while they’re in the car or play music from their phone or in this case they’re recording the video of the crash presumably because that will help Tesla understand what potentially went wrong and make sure that they’re improving the, the safety features on their vehicles. So accidents don’t happen in the future.
Lexy: Yeah. The article also spoke about the fact that Tesla has used these types of videos in other circumstances to prove that their autopilot feature was not responsible for crash incident. So this is as much a covering themselves as it is an ability to provide evidence for let’s say, investigators who are looking at how a crash occurred or what have you, if they receive this information back from the vehicle. Also to help improve their algorithms or help improve autopilot functions in the future for crash avoidance and so forth. Tesla has all
Marie: so built a feature into their vehicles that is called sensory mode, which basically allows the car to record what’s happening around it even when it’s not being driven as a way to help protect the car from either vandalism or from people that might try to steal it.
Lexy: Yeah. In addition, it has an interior facing camera system so that it can record who’s in the vehicle. If somebody does get in, that’s part of century mode also, so that means that if someone were to get in the vehicle and central mode was activated, you would actually see what they were doing in the car. If it was an nefarious actor who was entering the vehicle illegally, you could potentially identify who that was and provide video of what they were doing and so forth. I could see there being some good uses for that in other applications. If you think about ways that you could use that in AI. So let’s say for example, you wanted to develop an algorithm that could identify one a driver look tired and try to maybe engage them in something that would keep them alert or wake them up or what have you so that they were more attuned to what was going on on the road. You could potentially use an interior rear facing camera to do that. But we’re not seeing that type of technology yet. So having an interior facing camera just kind of makes me wonder how much of the drivers’ privacy are you taking away by having this camera or any of this other data being collected?
Marie: Well, and with the interior facing camera, that could also be something where they anticipate building that type of functionality in the future. And so if they know they want to be able to release it in the future, they might be building in the hardware for it now. So when they have the software can just be pushed or they could use it for detecting if somebody looks like they’re under the influence of something. And have similar safeguards in place that would make sure that the car could still be operated safely. Even if it looked like the person was under the influence of something
Lexy: or maybe it could just turn itself off and stay off if it thought that somebody was under the influence.
Marie: Sure. Potentially
Lexy: until a different driver came in.
Marie: So I think what’s interesting looking at this case is if you look at it from the perspective of Tesla, they are trying to anticipate adversaries, which could be people that are trying to steal a Tesla or trying to vandalize a Tesla or actual drivers that might not be in the best condition to operate a Tesla. And what can they do to build a safer vehicle that has greater safety features and also helps avoid crash risks. They have looked at developing these different types of systems to achieve those goals. But then on the flip side to do that, they are collecting a lot of information and are they taking the appropriate steps to protect the privacy of the information they’re protecting?
Lexy: I’m going to say it’s a hard no because they’re not taking adequate precautions. It’s one thing to say that we’re going to sync up to your phone so that you can have a good user experience with the vehicle. It’s another to say all of the information is going to be downloaded into the vehicle and stored in an unencrypted manner without your knowledge. So think of this from, again, from the perspective of this being a company vehicle, you as a company employee go into the vehicle, you sync up your phone, which is maybe a company phone. It has all your contacts, you’re driving around in it. You aren’t even the one that got the briefing on what the data was that was going to be collected from Tesla cause you weren’t the one buying the car. You’re just using it. What happens then if you’re renting a Tesla or something like that where you’re only temporarily using it but it still has all of your data and it’s not taking any sort of precaution to safeguard that information. It seems to me that that Tesla has not been as responsible as perhaps they should have been in protecting the privacy of users regardless of the fact that they have a very good reason for gathering some data. Other data really should be protected in a much more robust manner.
Marie: Absolutely. Kind of the flip side is that Tesla has said that there are options that people can use either access their data or to opt out of it, but it’s not giving users a fair choice because apparently in order to get access to the data, and it’s not even all of the data, you need to buy a cable that’s almost a thousand dollars and you can opt out of basically data being collected on you, but then you don’t get any software updates, which is one of the key features that Tesla owners benefit from, is that they’re driving a car and you’re connected to the system where you get your software updates and so you always have the latest and greatest features. So in order for them to really be offering a true choice, they would need to have an option that says you can still get software updates. Even if we’re not collecting personal that on you and there should be an easier way for people to review and or have some control over the data that is being collected.
Lexy: It really makes me wonder how Tesla gets away with this in the EU with Gdpr or how it will get away with this in California, come next year with the California Consumer Protection Act. Those regulations are very strict as to what data you collect and how it’s stored and what you can do with it and how you must remove it if requested to do so. And those types of questions I don’t think had been asked. And now that this is coming to light, I’m thinking they’re going to be asked. I just asked it on air. So if I’ve asked it, somebody else is thinking about very true. Somebody to whom it actually relates cause I’m not in either of those locations and you don’t have a Tesla and I don’t have a Tesla and I’m feeling pretty good about that right now. To be honest, I got to say as much as I’m a technophile I don’t want my car or my house or my devices to be that smart sometimes because this type of stuff worries me.
Marie: Having control over things make sense. And I do see the promise of electric vehicles and I do see the promise of some of the diff additional safety features that are being developed and we’ve even talked about in some of our previous episodes, the self driving cars and the technology that is being developed there. I believe we reviewed in those episodes the upsides, but we also talked about the pendulum downsides and what needs to be covered there.
Lexy: I think Tesla needs to take more responsibility and protecting the privacy of their customers and their customers contacts. And a lot of the data that is being collected regardless of the potential good uses, it still should be collected carefully and they need to protect the privacy of that data of the people that they’re collecting it on. And the,
Marie: the fact that this is now known a Tesla vehicle is now a bigger target. And so yeah, it is now an area where they have to anticipate adversaries.
Lexy: Well they should be anticipating adversaries anyway because they’re in the article at referenced a few other situations in which Tesla has been basically called out for lax security practices. One in which there was somebody who was able to get access to a cloud server cluster essentially and mine cryptocurrency using it. And it, oh by the way happened to have a bunch of autopilot data and another situation in which there was a wireless network list that had username and passwords in the clear. So this isn’t their first foray into some very murky waters around protecting privacy and having data security policies that were not sufficient.
Marie: And for a tech company developing this level of product, you would definitely expect there to be more consideration around these topics.
Lexy: Exactly. And that they would react much more quickly to ensure that they’re meeting the privacy needs. And I don’t see that happening. It talked about the fact that it took an extended amount of time for each of those other concerns to be dealt with. And that in this case, I mean this just came out in March, but it’ll be very interesting to see how long it takes them to and or rectify this situation with unencrypted data in their cars. Convenient that they have the wireless update capabilities to change that in their vehicles. But I wonder how much they’re actually going to do.
Marie: The other piece of it would be are there certain things that they have built into the system that they could continue to do, but just retain it for shorter periods of time?
Lexy: A lot of the data that they’re collecting, they want specifically for if there’s a crash. So if the vehicle does not detect a crash, maybe it drops everything. But the last 10 minutes, like on a rolling basis, the last 10 minutes of data that it had or
Marie: or it keeps the, the latest drive because maybe there’s something in the drive from the time it starts to the time it ends. Sure that would be helpful to review, but if you start the car and you end the car and there was no crash detected, then you don’t need to keep that data for our friends.
Lexy: Right. Or maybe when the car is turned off it removes all of the personally identifying information so maybe it would retain that. A call came in from a given number, but it wouldn’t associate that number with a name or an email address. All of that data would be removed as soon as it’s unpaired from the vehicle. That might be a way of removing some of the Pii data that could be really problematic that’s currently being stored in the car. There are a couple of different things that might be kind of a weird part of that is what happens when there is a crash. Obviously in this case these cars were salvage because they had been crashed. If the car crashes, what does it retain and then how is it safeguarded. That I think is the bigger concern is if it can’t automatically drop data or it needs to retain that information, how do you make sure that it’s retained in a way that does not grant access to anybody who happens, have the right kind of wire to hook up to the car
Marie: and like we’ve said before on these podcasts, we don’t always have the answers. Sometimes it’s just coming up with the questions and lots and lots of questions. When it comes to data science ethics, there are a lot of questions to consider, so that’s where we’re at with this one
Lexy: for now until there’s updates from Tesla.
Marie: It’ll be interesting to see how they respond to this over over the next few weeks. So that’s our quick take on the unencrypted information that has been found on Tesla model threes. Thank you so much for joining us for this episode of the Data Science Ethics Podcast. This is Marie Weber
Lexy: And Lexy Kassan.
Marie: Thanks so much.
Lexy: Catch you next time.
We hope you’ve enjoyed listening to this episode of the Data Science Ethics podcast. If you have, please like and subscribe via your favorite podcast App. Also, please consider supporting us for just $5 per month. You can help us deliver more and better content.
Join in the conversation at datascienceethics.com, or on Facebook and Twitter at @DSEthics where we’re discussing model behavior. See you next time.
This podcast is copyright Alexis Kassan. All rights reserved. Music for this podcast is by DJ Shahmoney. Find him on Soundcloud or YouTube as DJShahMoneyBeatz.