Autonomous Cars are about more than Driving.
With Tesla’s Model 3 officially starting delivery we should be thinking about what other data is going to come from autonomous cars. More importantly, what should we do with it?
True autonomous vehicles are coming in the next 5–7 years. Luxury automakers from BMW to Lincoln have been pioneering advanced driver aides and interventions for nearly a decade. Tesla turned on their controversial Autopilot beta program last year and Mercedes has promised us a fully autonomous vehicle by 2021. Eleven states (Alabama, California, Florida, Louisiana, Michigan, Nevada, North Dakota, Pennsylvania, Tennessee, Utah and Virginia) and the District of Columbia currently have legislation around the testing and use of autonomous vehicles. It’s now clearly a matter of when, not if, self-driving cars become the future.
As someone who loves to drive there are parts of this truth that bum me out and the idea of conditioning an entire generation of drivers to actually be less attentive on the roads is particularly troubling, much of this is cultural adoption challenges and will be overcome with time. For the rest of it — well there are still people who believe that vinyl is king so I’m sure a niche population will still be manually driving well into the future.
But as a technologist interested in the data around us I’m constantly reminded that autonomous cars mean more to society than fewer automobile crashes, improved quality of life for people with disabilities, and less congested roadways. Each autonomous car on the road brings with it a broad assortment of sensors including two of my favorites — LIDAR and Radar. LIDAR and Radar are indispensable in mapping spaces while creating virtual worlds. With IHS estimating nearly 76M autonomous vehicles on the market by 2035 that’s a lot of new data being created from nearly continuous scans of the routes of those vehicles. Take just a moment to imagine the possibilities.
Think about every municipality gaining access to a real-time virtual scan of their infrastructure. No more waiting for multiple reports of potholes, downed trees or powerlines to come into the local Department of Public Works before a repair crew is dispatched. A car driving by an impacted site can automatically alert the DPW that can then send out the necessary repair crew all while doing its task of taking you to work/home. Rapid repairs to infrastructure are a key to increasing infrastructure longevity. The benefit to mapping and navigational tools for those non-autonomous cars is also great. The increase in accuracy of applications like Waze when users no longer have to rely on others taking the time to manually tag an accident or road debris, or hidden police officer. If the cars could do it automatically we’d all get much more actionable data, both autonomous cars and non-autonomous cars.
By combining LIDAR scans with geo-positioning data it will be possible to get a closer-to-real-time view of the population changes in a number of environments. This data can then be used to ensure that adequate services are being provided to the community and help municipalities more accurately measure density (I’m looking at you MTA).
In addition, when coupled with machine vision technology the LIDAR and GPS data from autonomous vehicles can be used to monitor the opening and closing of stores and restaurants. Imagine knowing when you get in your car that the store you were planning on visiting had closed early, but that another comparable store 5 minutes away was still open. Drivers will have the ability to transform errands from “I have to go to X” to “I need to pick up Y” where you can go to the store that has what you need based a number of variables including if the store is still open.
Of course with any technology that’s in public spaces there are some very serious privacy concerns. LIDAR married to machine vision and GPS could potentially be used to effectively track individuals and groups. It could be used in this manner for good (think following terrorists and other bad actors) or bad (talk about giving conspiracy theorists something to work with). Remember when Germany decided to blur out all of their citizens in Google Maps? Well I think the reaction to this data would make that experience look quaint in comparison. I also believe there would be a similar reaction in pockets of the US as well, something that didn’t really happen with Google Street Maps.
However, much like in the 2010 Google Maps case in Germany only the publicly available result of the data would be obfuscated. The actual data the mapping companies would be selling likely wouldn’t be effected in the same manner. And this is a very important distinction to be made. Each car company is most likely going to have its own data set that it will be selling, possibly in conjunction with a technology partner. But in an industry that is slow to develop and extremely reluctant to adopt technology standards it’s hard to imagine the auto manufacturers ceding control. Just think about the fact that Apple Carplay and Android Auto were in the wild for over 2 years before they were implemented by automakers, and it’s still far from being a standard. I think it’s fair to say that automakers are not going this let this go to their partners. There’s also the fact that the auto industry generally follows a three year cycle for innovation and there’s little to no way that it’s ever going to be able to adopt the tech industry’s 6–18 month cycle.
So what we’re likely to end up with is different manufacturers with slightly different data sets that are all incentivized to sell that data to the highest bidder. The personal security implications are obvious — if there’s database of your every movement that’s for sale personal anonymity goes straight out the window. That’s not even mentioning the potential impact on civil liberties. Unfortunately the only thing that’s slower to react to technology than the auto industry is the government, which means that most of this will likely be hashed out in the courts.
But don’t forget this is merely one possible scenario. There’s always all of the good that can come from having all of that sensor data out there — even if most of it isn’t publicly available.
