Interesting self-driving design challenge involving signaling intent to outside world

As we get closer and closer to driverless cars, more questions regarding human-car interaction have arisen. One category of these discussions revolve around whether driverless cars should communicate with the pedestrians, bikers, and others around them to inform these people what the car is doing or planning on doing. For instance, if a car has stopped in front of a pedestrian, maybe it’s waiting for the pedestrian to first cross. Does the pedestrian feel confident that’s why the car has stopped and would this person feel comfortable crossing?

Some companies like drive.ai are tackling this problem by incorporating digital signs on the outside of cars to inform the outside world of what the car is “thinking”. Others believe that self-driving cars should mimic the conventions of normal cars such that people already have a set of conventions that they can continue to stick to. For the former approach, there are a lot of challenges related to universal design, or making sure design works for all kinds of people. If the car shows English words on the outside of the car to communicate intent, what about the people who only can read Spanish or Mandarin? How about people with impaired vision? If the car shows symbols or pictures of its intent, how big of a vocabulary can this produce to capture all the different intents? Can people easily learn these meanings fast enough for this to be useful? Would these explicit signals give people a false sense of confidence such that they don’t pay attention to traditional signals of the car? For the approach of mimicking the conventions of current normal cars driven by people, can we expect for bikers or pedestrians to trust that the historical conventions will be followed by these new foreign car systems, or will they assume that these cars are ignorant of those conventions?