10 Reasons You Won't See The Cadillac InnerSpace On The Road Anytime Soon

Perhaps the biggest issue with the idea of a fully-self driving car is the limitations of current autonomous technology. While systems like Tesla's Autopilot and Cadillac's own Super Cruise are mostly capable of driving down highways without human intervention, come away from the main roads and they're not so capable. American cities with grid-like infrastructure are relatively easy for developers to map, and most urban-focused development work has taken place in these environments. Google's Waymo division, for example, launched its first autonomous driving service in Phoenix, Arizona, while Uber's development program (and its fatal crash) took place in nearby Tempe, alongside a separate testing program in California.

However, look at cities in other parts of the world — notably, Europe and Asia — and cities tend to be more cramped, unpredictable, and difficult to navigate. If self-driving cars are to become viable, they'll need to be able to navigate these packed urban environments with near-100% precision, and they simply can't do that right now. Things get even trickier when you leave the city and get out into rural areas, as varied terrain, limited connectivity, and a whole new set of hazards present themselves. While today's best autonomous systems might be good at dealing with a limited set of circumstances, they're nowhere near being able to deal with everything that real-world driving might throw at them.

Even if autonomous systems can reach the point where they can comfortably navigate through all of these situations, there's the additional problem of proving to regulators that they can actually do what they claim to. During the first few years of autonomous car testing, there were plenty of authorities willing to give the technology a chance. However, Uber's 2018 accident played a big part in dampening that initial wave of enthusiasm.

A Volvo XC90 fitted with Uber's autonomous driving system plowed into a pedestrian who stepped out into the road unexpectedly, leaving her with injuries so severe that she died after being taken to hospital. Analysis of the crash showed that Uber's system had detected her presence, but deemed her to be something that the car should drive through rather than attempt to stop for. Essentially, the car decided to crash straight into her. It's worth noting that, in this case, the car did have a safety driver who was distracted from the road, and she was eventually charged with negligent homicide. However, any fully autonomous car, like the InnerSpace, will have no human input at all. So, the system will have to be trusted to make all the decisions by itself.

Even if, and it's a big if, GM did suddenly decide to sign off on a program to make self-driving cars a reality, the cost of development would make finding a suitable retail price for these cars an almost impossible task. To get anywhere close to recouping its costs, GM would need to set prices much higher than any equivalent non-autonomous luxury car, Cadillac or not. That would put the production InnerSpace in a price bracket all of its own, competing against cars that it likely couldn't match in craftsmanship or prestige.

It's also not clear what segment the InnerSpace would aim to compete in, for example, whether it would aim to attract buyers of the Mercedes S-Class, or position itself even more upmarket to take on the likes of Bentley. Targeting either one would require a build quality and finish far superior to that of the average GM product, and for that matter, considerably better than any current Cadillac on the market. Even then, the added cost of developing full autonomy would mean that GM would likely be working on tighter profit margins than either the German or British manufacturer. The concept itself might be new and innovative, but that innovation comes at a cost, one that will have to be directly reflected in its MSRP.

While the InnerSpace's central sofa-style seat is great for occupants to make use of the large curved screen fitted to the front of the cabin, it does pose a greater risk than a traditionally-styled car seat. With passengers able to sit very close together and with little securing them to their seats, there's a chance that they could collide with each other in the case of an accident. This doesn't matter in a flashy concept, of course, but it's something that Cadillac's design team would have to consider before reaching the production stage. Modern seat belts are designed the way that they are for good reason, and the seat's shape and maneuverability would make it tricky to fit an effective one.

Another impractical feature is the doors and roof, which open both upwards and outwards when the occupants get out of the car. That setup leaves it vulnerable to rain or snow getting in, and might even become non-functional in places where heavy snowfall is a regular occurrence.

Assuming self-driving cars can pass all of these hurdles and get onto public roads in the first place, there's still the legal issue of whose fault it is if said self-driving car is involved in an accident. Let's say an InnerSpace, which has no way for the occupants to take control at any time, hits a pedestrian and kills them (like Uber did in 2018). Who bears the responsibility for the crash? The car's occupants will be focused on enjoying their "wellness experiences" if Cadillac's press release is anything to go by, so they won't be looking at the road. Even if they were, there's no way for them to stop the car. Surely then, it can't be their fault if an accident happens.

There are two other main parties that could shoulder the blame: insurers or Cadillac itself. Cadillac certainly isn't going to want to take on that responsibility on its own, so that's out of the question. Whether insurers are willing to take on the risk of paying out for the actions of an AI system remains to be seen, but let's face it, it's costly enough trying to get them to insure humans. It's arguable that an autonomous fleet of cars would eventually be safer, but this assumes their systems will be vastly more capable than today. Plus, as long as there are unpredictable human drivers on the road, there will still be plenty of accidents (and as a result, costly lawsuits) to deal with.