Making Money On Self-Driving Cars: The Roving Eye Will Be Golden

At many of my speaking engagements on self-driving driverless autonomous cars, beyond the usual technology-based questions, I also oftentimes get asked various business-oriented questions about whether or not autonomous cars will be profitable.

You might have already assumed that self-driving cars will be profit-making vehicles.

Indeed, it is certainly a natural assumption to make. Why in the world would all of these automakers and tech firms be toiling away at trying to make, build, test, and deploy driverless cars unless they felt pretty strongly that there was a decent profit to be made?

In fact, some are suggesting that whoever gets to the vaunted true autonomous car first, the Level 5 which will have an onboard AI system driver that can take you anyplace that a human driver could, will be in the bucks, they’ll get the reward of vast riches for their labors, and be soaking in obscene amounts of money.

Some AI developers are slogging forward because the Level 5 is a moonshot-like challenge and offers a puzzle to be solved unlike any they might otherwise encounter (in some respects, the challenge being more important to them than the money aspects per se). Some are interested in the acclaim from being able to fully execute the “mother of all AI projects” as proclaimed by Apple CEO Tim Cook. Some hope that achieving an autonomous car will democratize mobility, freeing people to get around, and revolutionize how our society functions.

Some want the money, or possibly a combination of the aforementioned loftier aspirations plus the money.

Ways We Make Money From Cars

The most obvious ways that everyone imagines that there is profit to be had will be these traditional kinds of modes:

• Selling autonomous cars

• Renting out autonomous cars

• Delivering via autonomous cars

• Ridesharing out autonomous cars

If you ponder those key modes, it really doesn’t seem to be much different than today’s world and contemporary conventional cars.

You can make a profit by selling conventional cars.

You can make a profit by renting out conventional cars, though admittedly the car rental agencies are a bit queasy about a future of driverless cars and how that will mess with the renting of cars.

There’s also the use of cars to do deliveries, such as the current craze of getting a taco and burrito delivered to your door at midnight, which many are already experimenting with via using autonomous vehicles lesser than full-on cars to do this kind of simpler and more bounded act.

You can maybe make money by ridesharing out conventional cars, but the record so far of Uber and Lyft doesn’t give you any warm and fuzzies that a real profit is to be readily had (they are losing money hand-over-fist, though some say this is just the early part of the service life cycle when trying to grab market share and gain a loyal base, so you need to expect that sometimes you have to spend money to ultimately later on make money).

For ridesharing of conventional cars, the difficulty of trying to eek out a profit involves all of the costs associated with the ridesharing arrangements. Pundits are expecting that once the human driver of the ridesharing car is excised from the deal, which presumably will take place once autonomous cars are safe and available, the most significant cost that has been a presumed major impediment toward profitability will be off-the-table (i.e., human driver labor).

Smooth sailing afterward, they assume. The thing is that we don’t really know yet what the financial numbers will really look like. What will be the cost of the AI system and its ongoing updates and maintenance? Will there need to be special kind of insurance for driverless cars and how costly will it be? And so on.

Let’s though put aside the usual modes of making money from a car, whether it be autonomous or not, and use our thinking caps to find other ways to make money from a car.

Other Ways To Make Money From Cars

I’ll give you a clue about another likely significant way to make money via a car, especially down-the-road in an era of driverless cars (that’s a teaser, a spoiler alert of what I’m about to reveal!).

Recently, ridesharing drivers have been selling their eyeballs, well, the use of their eyeballs, by contracting with real estate firms that are trying to find houses to flip.

If you are an Uber or Lyft driver, you are presumably roving around for hours on end, hopefully with paying passengers on-board, most of the time, and have not much else to do other than drive the car. I realize that human drivers are supposed to be fully attentive to the roadway and traffic, but I think we all know that humans aren’t nearly that singularly focused when behind the wheel.

So, a ridesharing driver can easily keep their eyes peeled for any houses that seem to be viable house flippers.

Usually, telltale signs are when a house looks like it is pretty much semi-abandoned, including that the front lawn is disheveled, there might be posted signs warning that the property is in violation of homeowner’s ordinances, and so on. It’s pretty easy for a ridesharing driver to be watchful for such houses. Once they spot a potential house flipping candidate, they either write down what they spotted, or enter it into a mobile app, and they’ve just completed a tidy extra task.

Along with the tidy extra task comes the potential for getting a tidy fee from the real estate firm seeking houses to be flipped.

In short, another way to make money and possibly profit from a car would be to use the car as a handy platform from which to traverse an area and find or detect something that others would be willing to pay to know about.

Fortunately, autonomous cars are going to be able to do this in spades, escalating exponentially the capability to rove and scan, it’s in their bloodline, one might say.

Driverless Cars As Roving Golden Eyeballs

Imagine that there are autonomous cars undertaking ridesharing.

Some assert that this is going to be a springboard towards ridesharing on a volume and scale that we can hardly conceive of. If that’s the case, there are going to be driverless cars roaming and roving all over the place, taking human passengers to their destinations, picking up human passengers that need a ride, and otherwise trying to be in the right places at the right times.

For purposes of being a driverless car, the vehicle is jam-packed with sensors. The sensors include multiple cameras, collecting visual video and images that are being interpreted by the AI to figure out where the road is, where to turn, whether pedestrians are nearby, etc. Plus, there are other sensors including likely radar, ultrasonic units, Lidar, thermal devices, audio listening devices, and a cornucopia of similar sensors.

The mainstay at first will be that those sensors are there for the purpose of driving the car. Don’t mess with the sensors or the onboard computer processors for anything besides making sure the autonomous car can safely get from point A to point B. I’m betting that there will be some extra computational cycles available, perhaps normally kept in reserve for situations of a tight or tough driving situation, and yet otherwise could be used for other aspects if needed, when viable to leverage.

Why not have the AI look for houses to be flipped?

It would be relatively easy to do.

The data pouring in from the cameras is already going to be undergoing analyses. You could either have the vision processing system tag the images of houses and in real-time be marking ones that could be house flippers, or, you could set aside for the moment the house flipper search and have that undertaken in more idle moments.

For driverless cars, there will be idle moments for example when sitting at a red light. You and I likely daydream when at a red light, but the computer processers on-board can be doing some real work, like rescanning images to find those house flippers. Or, maybe when the driverless car is getting charged-up at an EV charging station (most self-driving cars are likely to be EV’s), during that idle time the processors can be doing constructive acts such as filtering the collected data for houses meeting the house flipping criteria.

Let’s kick this up a notch.

Fleets of driverless cars will be likely uploading their individually collected data to the cloud database of the automaker or tech firm (or other). The uploaded data might already have been examined for houses to be flipped, or this might be a task undertaken once the data has been pushed up into the cloud (sparing the on-board AI systems this task).

The collective “wisdom” of hundreds or say thousands of driverless cars that have been whisking around towns and cities, well, it’s a treasure trove of handy data.

What Will The Eyed Data Be Used For

Think about the possibilities.

You are a local house painter and you are trying to find out which homes in the county need a paint job.

Easy, just consult a database of houses that have been “seen” by a fleet of driverless cars.

You can search the database by where the house is, when it was most recently seen, and pull-up an image or two (or a video) that shows what it looks like. The AI can already pre-assess which homes have faded or peeling paint, so you don’t need to cull the list.

This can be tied to another database of homeowner records, and with the touch of a button you can be on the phone to the homeowner, explaining how wonderful their house would look if painted in green. You can even text them a photoshopped version of their house with the color changed.

Right now, we often take a look at a large database such as the one maintained by Google, trying to see what a house looks like. Those images are often dated, and there are only so many times that they are able to send out their handful of photo-taking cars to capture an area. With driverless cars, assuming they are already on a ridesharing quest, this photo-taking or video-taking is part-and-parcel of what they are already doing. No added cost, in a manner of speaking.

It is conceivable that via various ways to make money from the driverless car collected data, it might motivate the autonomous car owner to keep their self-driving car on-the-go, even when there aren’t any paying passengers needing a ride.

If the money is good enough, it might be better to actually avoid taking rides, since those passengers expect to get to their desired destination, which might not be where your driverless car needs to go to collect money-making data.

Downsides Of the Madcap Data Collecting

The world is never accommodating to having only benefits and no costs, or so it seems.

In the case of the driverless car data collection, once a particular neighborhood has been scanned, the question arises about what good does it do to potentially scan it again a few hours later. In essence, it could be that the data won’t have quite the payoff at all times as might be hoped for.

Plus, if one fleet has already done that neighborhood, and a different fleet does so, this will rachet up competition in terms of these collected databases, meaning that there might not be as much profit as one would have if only one fleet had the golden goose and no others did.

The privacy implications of this kind of vast and unending data collection can be quite staggering when you sit down and think about it. I’ve only mentioned the notion of scanning to find houses, but it is perhaps obvious that the scanning of the images can look for many more objects, including human beings in the images.

Want to know where your significant other was yesterday? Log into a cloud database of a fleet of driverless cars, give it a picture of your partner, and let it find them, reporting to you that the person was seen in front of the pancake restaurant, then later on seen at the local park, then at the door of an apartment in downtown, and so on.

Conclusion

The main theme of this discussion was to consider ways to make money from autonomous cars, beyond the obvious means that we do today with conventional cars. A key difference of conventional cars and driverless cars is that you’ll have a full suite of sensors, collecting data wherever the self-driving car goes, and the data can be kept and analyzed on-board the driverless car or placed into the cloud and used there.

This can be done in a scale and in a manner not feasible for what human drivers could do. There are likely money-making ways to leverage this that no one has yet envisioned and will emerge once self-driving cars become prevalent. Try to figure that out, before everyone else does.