One "story of the year" for Time was the #metoo campaign, where (mostly) women shared stories of how they had been sexually harassed or molested, to make it clear just how widespread the problem is. Almost all women have a story, or many stories, sad to say.

I have frequently heard reports from women of being groped on crowded public transit. People are packed in, and villains use the plausible deniability and anonymity of the packed crowd to grope.

This was spelled out in a Washington Post story on metoo and transit (paywall, but I suspect everybody knows its truth.)

This adds a new positive element to the end of conventional public transit as it switches from large, human driven vehicles on fixed routes and schedules to ad-hoc use of vans and cars with robotic driving -- the question of public safety, and not just against groping.

The first reaction is often the other way. Transit vehicles have a driver on them, and even have transit police. This keeps the public in line. Even an UberPool or Lyftline, where you get into a car with a random stranger, still has the driver there who can assist if there's a problem between passengers. In a robotic shared vehicle, it might be just a few people, or even just two, and nobody else. On an empty road at night. On the other hand, there is much less motive to share robotaxis at night, when the fleet is mostly idle, roads are clear and trips have less in common.

I've spoken to women who do not feel safe in such a situation, and would not ride in such a vehicle, at least with an unknown man. They might wish for the ability to request this in their ride profile. (I have discomfort, though, with institutionalizing the enablement of discimination based on sex, race or other inherent traits.)

One answer for that is that services like UberPool are not anonymous. Because Uber knows who everybody in the carpool is, that's a major deterrent to bad behaviour. A bad actor would not just be kicked out of Uber, but possibly reported to police. Uber knows even more about who its drivers are, but that hasn't stopped there from being fear (mostly promoted by taxi drivers who don't like Uber) of attack by an Uber driver.

Identification sounds good, but it's also a privacy nightmare. Uber requires a credit card and so is unavailable to a large swath of society who can't get such cards. Public transit and taxis have offered us a privacy-positive world of anonymous transportation. Many of us feel that it is a fundamental right to travel anonymously, without being tracked. (Though so many of us have given that up to our cell phone provider.)

How do we reconcile the needs for private travel, safety, the feeling of safety and access for the poor?

Compartments

One approach, possible only in small to medium vehicles, is a shared car or van with private compartments. A shared taxi might be designed to have 4 doors and 4 compartments with a wall down the middle so that each person is secure and private. You can't design a regular car that way, but a purpose-specific taxi could be built with that design. You could go all the way up to a van. My 16-person van design (see link above) could have 8 compartments if it had extra doors.

For crash-safety, however, it must be possible if the vehicle rolls on its side to get out the other side. This requires there be a means to get through the divider in an emergency, and without much difficulty. This adds a complication to the problem.

Another creepy situation could arise if a person gets out at a stop and the person in the next compartment says, "I've changed my mind, I will get out here too." (You can't imprison people in the car so this must always be an option.) Of course this has always been an even greater risk on public transit which has complete anonymity.

Preferences

People could express preferences on shared rides. Outside of rush hour, a private ride will always be available, though it may come at an extra cost. During rush hour a private ride may be much more expensive, but crowds are larger and they provide protection. Those with fear -- and money -- can solve their concern that way.

It is possible that the system might allow a woman to say, "I will share only with other women" though it might mean a slight delay in allocating a vehicle. Or they could say "I will only share with at least 4 at all times."

Society might tolerate such a request, but it would likely not allow riders to say, "I will only ride with other white people" and would have a difficult debate over what's allowed and what is not.

If we allow anonymous riding -- and we should -- it might allow some constraints on that. You might insist on riding only with identified people. In this case, you might have to pay more for that privilege, otherwise private travel would be seroiusly impeded. A better solution might have the identities of riders held by a trusted escrow service, which would erase the travel records after the successful completion of the ride. The car could also photograph anonymous passengers, again ideally destroying records a day after the ride, and only offering them to police with a warrant.

Even with identity escrow, there is a lot of privacy risk. Today's world of public transit is fairly anonymous. You can put cash into a box and ride with nobody knowing anything about it, though train platforms often record video. This world could easily end in the world of robotaxis or even of Uber.

No crowds

But even without private compartments, or identification, or preferences, robocar transit is going to be different. The packed bus or subway car will either vanish, or become an option easily bypassed without a lot of money. And that means, at least, the end of the random groper on the train.