Law & Technology on the Bleeding Edge: The Future of Google's Automated Cars

Another great guest post from TommyLoff. This time he's flexing his legal muscle.


- 44 Maagnum




Thanks to the fine people at Carpe Daemon (and perhaps a few other small news outlets) our readership (potentially 7 billion strong!) is fully aware of just how awesome Google’s automated car project is. Very. In the past six years, autonomous vehicles have gone from completing 7 of 150 miles in the 2005 DARPA Grand Challenge to an incident free 140,000 miles on California highways by a set of Google-enhanced Prius’.  News of this relatively secret project broke in the fall of 2010, and Google has since busied itself with a few nifty demos and lobbying efforts to establish the legal private and commercial use of their technology in the state of Nevada. The question, “Why Nevada?” remains unanswered but the notion of self-driving cars burning rubber down the Las Vegas strip is a world I very much want to live in.

Google’s lobbying has culminated in the presentation of two bills to the Nevada state legislature. The first would allow for the licensing and legal operation of autonomous vehicles on Nevada roads, and the second would exempt the operators of said vehicles from the statewide prohibition on texting while driving. These bills are likely to come to a vote in the next week before Nevada’s current legislative session comes to an end. Passage of these bills is certainly an exciting premise for tech enthusiasts and futurists alike; however danger lies in whether or not our legal system can cope with such rapid technological development.

Such legislation is based on Google’s assertion that self-driving cars will not only be safer, but also dramatically reduce infrastructure costs in our transit system by decreasing traffic delays and increasing fuel efficiency. These promises would result in an annual aggregate savings of well into the billions as well as hold the potential to save hundreds of thousands of lives annually. Gains however are not without costs, and the reality of yielding personal autonomy to lines of code will likely require a dramatic shift in our existing legal paradigms - particularly how we consider liability as a society.

Thankfully such a change will not happen with the flip of a switch. There will be a significant period in which humans share the road with a small minority of autonomous vehicles and only slowly, over many years, as the number of autonomous vehicles increases will the societal level benefits begin to become evident. Let’s begin with this transition period - one which we may officially be entering within days depending on the decision of the Nevada legislature. At the outset, liability will continue to fall on the vehicle operator with exceptional circumstances excluded. If somehow for instance the automated software prevented the driver from effectively controlling the vehicle thereby leading to a collision then perhaps product liability would come into play. For the most part however, existing doctrines of contributory and comparative negligence, which establish the criteria necessary to seek compensation in a civil suit, will continue to be used to assign liability to vehicle operators. (For the legal eagles among the readership, contributory negligence prevents compensation in a civil case if the Plaintiff is deemed to be even 1% liable. Comparative negligence attempts to establish what percent each party contributed to an accident and base settlements off of these figures.) While liability will continue to fall on the operator for the foreseeable future, legislation will continue to narrow the scope of this liability and laws governing the usage of automated technology would develop piecemeal on a state by state basis to reduce the culpability of individuals as they transition from vehicle operators to mere occupants. With the Nevada bill already pushing for an exemption from the state prohibition on texting for automated vehicle operators, it is clear that the envisioned end game is a system composed predominantly of automated vehicles in which vehicle occupants are divorced entirely from the vehicle’s operation.

This is where things fall apart. When the vehicle occupant loses culpability for the action of the vehicle, a new paradigm must emerge; particularly if we are operating under the assumption that automated cars still collide from time to time. Regardless of how dramatic a safety improvement we may see, the idea that automation can bring our annual highway death toll to zero is extremely far-fetched. Currently the highway death toll in the United States is roughly 400,000 annually. As a thought experiment let’s consider that a future system, with negligible human drivers, reduces this number to 100,000 deaths annually. Outwardly this is a very positive figure – yet one which lands us in a pickle. The only potentially liable party for motor vehicle accidents is now the product responsible for improving aggregate safety by 75%.  Unfortunately we are a society which thrives on the assignment of blame. If Google and other manufacturer’s automated systems are reducing American highway fatalities from 400,000 a year to 100,000 a year is it truly reasonable for them to bear the burden of 100,000 wrongful death claims despite preventing an additional 300,000? From both an ethical and practical standpoint it would appear that one must side Google. Such dramatic safety improvements should be rewarded not punished, and as a practical matter, if risk is no longer dispersed across the hundreds of millions of motor vehicle operators in the U.S. and instead consolidated onto a handful of corporations, the necessary insurance premiums for these corporations could threaten the existence of the entire industry. Yet the idea that wrongful deaths would be written off as a necessary evil of our new and improved transit system is preposterous. As troubling as financial compensation for loss of life or well-being is, even more troubling is no financial compensation and I for one am somewhat at at loss for a solution. Luckily our courts and legislatures will have at least some time and intermediary steps prior to reaching a new equilibrium.

Obviously there is a long road ahead for the implementation of automated vehicles and there are tremendous obstacles ahead. For one, as demonstrated by Will Smith in I, Robot, people really don’t like the idea of robots doing people things; especially when it involves giving up personal autonomy. We are irrational beings, and just as the vast majority self-assess themselves as above average drivers so too will many cling to the belief that in a dangerous situation they would prefer to be behind the wheel as opposed to a machine. Thus it is unclear whether or not the ‘cool factor’ alone will be enough to drive initial implementation. That said in recent years it has not been wise to bet against technological advancement, nor against Google. Their demos, combined with the seriousness of their efforts towards bringing their product to market in Nevada and the sheer size of the market they could potentially capture speak to the legitimacy of Google’s operation. That there is dramatic uncertainty regarding the legal ramifications of such technology should be no barrier towards progress and concerns that we do not have a framework in place to equitably resolve every potential outcome ought not be used as grounds to limit progress. This is particularly true in this situation as the exact changes necessary to our legal framework won’t become evident until some level of implementation is achieved. Stepping beyond the boundaries of the autonomous car, the point at which technology and law meet on the bleeding edge must be approached with flexibility in mind. As we move forward, it is the adaptability not the predictive value of the law with which our legislatures and courts should concern themselves.