Self-Driving Taxis in San Francisco and Austin: Safety, Responsibility, and Personal Experience

Self-driving taxis, also referred to as “robotaxis,” no longer seem very distant or futuristic concepts either. In cities like San Francisco and Austin, self-driving cars have already made an appearance on roads without anyone at the wheel. Companies named Waymo have also begun charging for these “trips” in these cities, depending on sensors that include lidar and cameras, among many more, that help self-driving cars move on roads that were, at first, quite challenging for them. 

On the safety side, self-driving taxis have a mixed but encouraging trend. According to the latest safety reports by Waymo, their self-driving cars have been proven to be far less involved in severe crashes than their human drivers. For example, in the case of San Francisco, Waymo's self-driving cars recorded 91% fewer severe crashes involving injury or worse per million miles compared to the human baseline. The same is observed in Austin as well as other operational regions, especially in terms of crashes involving injury and airbag deployments. 

The Austin-specific data supports the safety advantage: Waymo's robot taxis in Austin experienced 81% fewer crashes resulting in injury and 94% fewer airbag-related crashes than typical human drivers. Overall, the trends look rosy, but there are some warts in the Austin data that include robot taxis illegally making passes at stopped school buses and some erratic driving, for which Waymo has issued software patches. 

However, not all autonomous systems are similar in terms of their level of safety. From the tests done by Tesla on its robotaxi on the road in Austin, crashes are recorded after each 40,000 miles, against the 500,000 miles recorded for human rates. These values indicate that not all autonomy is of an equal level of reliability. 

While it’s easy to simply state that it’s about calculating numbers, the real problem relates to determining who can be held accountable in case something goes awry. With human drivers behind the wheel, insurance policies along with traffic regulations provide a fairly straightforward way of determining who’s at fault in the event of an accident. However, when it comes to robot taxi accidents, it becomes difficult to pinpoint who should be held responsible – the developer of the algorithm, the sensor developer, or the government for allowing self-driving vehicles on public roads in the first place. Therefore, there are no answers as of yet, but it’s being discussed by lawmakers who need to find a way to hold the appropriate individuals responsible. 

These are some of the open questions in the area of law that fuel public uncertainty. A drive in San Francisco made me realize how fluid it was—the vehicle gliding smoothly through congested streets and intersection points. At the same time, there was an unnerving feeling of it momentarily holding back or braking too quickly in anticipation of some unexpected encumbrance that made me question its response to hard-and-fast situations. This is reflected in the overall anxiety about handing too much control to automation in situations that are demonstrably statistically safer. 

Nonetheless, the potential benefits of many robotaxi deployments being highly significant, they could reduce traffic accident fatalities and injuries significantly, since it has been observed that most accidents on roads are caused by human error.They could improve access to transport for those who cannot drive, whether by disability or mere old age, and reduce traffic congestion by optimizing roads and routes. 

Thus, self-driving taxis in urban areas such as San Francisco and Austin are a real-life preview of what the future of transportation might hold. There are initial statistics that reveal reduced rates of accidents compared to human chauffeurs, but this performance also depends on the company involved, and questions exist regarding liability for accidents. Personal experiences of the sort that I, for instance, share are but a reflection of the collective caution that the general public shares.