Autonomous cars are just now barely starting to end up on our roads, in limited numbers and in limited ways. For now. We’re in an interesting transitional period, and there’s still lots of things we have to figure out. I think I have an idea that could help, and it’s as simple as a dumb old light.

Right at this moment, there’s at least a couple of cars available now or soon capable of some degree of fully autonomous driving: the Tesla Model S and the Mercedes S 500 come to mind. There’s many other cars with near-autonomous features like active cruise control, lane keeping, emergency self-braking, etc.

These are just the start, of course, and the one thing they all have in common is that, from outside the car, you would have no idea if it’s a human or a robot running the show. I think this distinction is actually important, as is the ability to see and track the cars driving themselves.

This may sound a bit like I’m paranoid about a massive robo-car uprising, but that’s not it at all. I think being able to visually see where the autonomous cars are will actually help their acceptance into mainstream traffic, and provide tools to study their interactions with other autonomous cars, meat-driven cars, pedestrians, and all the other chaos of life on the roads.

Here’s what I’m thinking: any car that has some level of autonomous capability – let’s say at the level of a Tesla Model S or better – would be required to have a roof-mounted light. This light should be visible from 360° around the car, and, significantly, from above. I’m thinking the color of the light should be something not already in common use, which is trickier than you’d think.

Red, orange, and clear are all taken by regular car indicator lights; blue is for cops, yellow is often construction equipment, which pretty much just leaves green and purple. Green is actually sometimes used by private security, and is, of course, the ‘go’ color on traffic lights, so until someone figures out how to make a viable brown light, I guess we’ll have to go with purple.

I know, I know. But even if it’s purple, I think there’s still good reasons to have this.

The light would only be activated when the car is being driven autonomously. Distinguishing which cars are autonomous or not openly, obviously, and easily is going to provide several advantages.

First, I think it’s going to help smooth the process of acceptance of autonomous cars into traffic because people who are uncertain or don’t quite trust the cars will have the means to spot them easily, and avoid them if they choose.

It would be a gesture of openness, almost a respectful gesture to people who still drive their own cars. In the early stages of development, when autonomous cars really have yet to prove themselves en masse, I think drivers will be more open and willing to allow them to share the roads if they make their presence clear.

This can also help on the opposite side of things, too. For those who do genuinely believe in the safety of autonomous cars, and, even, their superiority over human drivers, those purple lights could be a beacon of perceived safety for some drivers.

And, speaking of safety, drivers of emergency vehicles like ambulances or fire engines could use the lights to identify where autonomous cars are on a road, and, provided we come up with a standardized procedure for what autonomous cars should to when an emergency vehicle approaches, those drivers could use the autonomous cars as a predictable path to their goal.

An ambulance driver approaching an autonomous car can count on a set series of actions for that car to take, something they can’t guarantee when barreling at some rando in a rusty Cressida. This could help those vehicles get through traffic more quickly and safely.

The reason the light should be visible from above is for easy tracking of autonomous cars, and their patterns. I’m sure manufacturers and researchers can get GPS and other data from their cars to study how they’re doing in various driving and traffic situations. That does not mean those manufacturers have to share that data with anyone else.

The overhead-visible autonomy light would allow independent researchers to study the behaviors and habits of autonomous cars without being beholden to the manufacturers. Cameras mounted to buildings watching city streets or drone-based cameras could look at traffic and see what cars were robo-driven, and how they behaved.

There will likely be very valuable data to be gathered by watching how the cars work in reality, and having a means to get this data independently of any on-car logging or anything like that could prove very useful.

Police could use it in car chases. You could tell at a distance what sort of driver your Uber had from a distance, meat or metal, and decide if you want one or the other. You’d know if a vintage car had been converted to autonomous control with just a glance.

Also, depending on how the laws eventually shake out, it could prepare other drivers and/or law enforcement to know what to expect in a given car. It may become legal to have a totally unmanned car, or a car with no driver, but carrying someone’s kids or dogs or elderly mother. A purple light will let people know there may not be a driving-age human adult in the vehicle.

Up until now, a car’s driver could really only be one thing: a human being (exceptions made for sassy orangutans). Now that there’s another option, it’s simply a matter of acknowledging, without judgement, that what’s behind the wheel of a car is valuable information to convey.

I’m sure auto designers will have some fun designing these into cars. There’s no reason they have to look like crap.

Contact the author at .