By: Gadi Perl.
I’d like to discuss the unique difficulties trying to regulate autonomous cars. Autonomous cars (more exactly Self Driving Vehicles, SDVs) are a technological advancement with great potential to change our lives. This change is much more than just the way we get from one point to another.
Let’s enumerate some of these changes for a moment. With a predicted fall in accident rates and a decline in the severity of injuries, the insurance market will surely have to adapt. Employment of driving related positions – truck drivers, taxi cab drivers, EMT crews, will change dramatically. Labour laws regulating working hours of drivers will become obsolete and other regulation regarding intoxication, sleeping hours etc. will require adaptation. Liability regarding accidents involving SDVs is a hot topic, since the driver, which has been the focal point of most tort related discussions – will become a machine – changing the way we decide who is actually to blame when bad things happen. Newer business models, which will decrease car ownership dramatically, will impact city planning, lowering the need for parking spaces near homes and businesses. Efficient driving is predicted to lower congestion allowing for savings in infrastructure, and lower costs in mechanical repair due to efficient driving. The potential of the future feels limitless.
It is not surprising, therefore, that public interest in SDVs is high with every day people becoming expert readers regarding Artificial Intelligence. The problem is that the endless potential of SDVs, combined with public interest, raises the stakes for legal scholars dealing with the regulation of autonomous cars.
This regulation task is quite complex. We need to a) regulate SDVs in such a manner that will simultaneously prevent antiquated laws from creating market failures, b) create new mechanisms to solve man vs. machine issues from the new technologies – and all this while c) creating public trust in the new technology.
AI driven vehicles raise psychological fears about machines controlling the world. Terminator movie series fears, about machines taking over the world after an AI named "Skynet" achieves consciousness and starts killing people – suddenly become closer to reality.
Most people will not actually be afraid from being killed by a machine and some may even embrace AIs and SDV. Nevertheless, I believe many will be hesitant of this new technology and the way it will affect their lives. The more moderate will refrain from buying SDVs or using them, and the less moderate may take action to limit the use of SDVs. This is already evident with some even intentionally crashing into SDV test cars on California Roads.
When people kill other people in car accidents, it does not appear in media headlines unless an unusual human interest story is involved. On the other hand, when an autonomous car kills a pedestrian, as recently occurred in Arizona – it almost immediately becomes world news. Every car accident involving an SDV is news worthy.
When it comes to SDVs the public has yet to deal with the impact of issues such as privacy, and their reaction upon realization of the risks is an unknown factor. How will the public react when cars with the potential of filming everything around them at all hours of the day will be in every corner? What will happen when the first SDV is hacked, sending all the passengers plunging to their deaths? The risks are high. This true especially when you take into consideration the vast amount of variables related to self-driving cars. A wrong decision by regulators following public outcry may impact the market detrimentally.
For example, a decision to force human supervision of SDVs at all times, caused by fear or by lobbying groups – can impact the rights of the populations such as the blind, who will not benefit this new technology, and may potentially discriminate the elderly drivers who may respond slower to incidents.
We must work harder to identify in advance the issues before they come to life. We cannot leave loopholes since any major fault not dealt with in advance may delay the proliferation of SDVs and will eventually cost us in human lives.
Legal Scholars Learning to Speak “Tech”
So after I have shown that this is a difficult task, let’s talk about the solution of technology oriented legal discourse.
It seems that when debating technology, being just a legal scholar isn’t enough. The new high-tech products currently entering commercial use are too complex for regulation without a deeper understanding regarding the technologies involved. If we don’t understand the mechanism required behind some of the SDV technologies, we will not be able to identify legal issues in advance.
A good example for the need to understand the technology is the cameras currently mounted on SDVs already being tested on California roads. These cameras have the capability of taking high definition in 360 degrees, constantly. These ever-documenting cameras raise many legal question which require technical understanding for them to be properly stated. Do SDVs require all the data from these cameras? If they do, do we require the data to be saved for all times? If so, how do we protect the data collected from these cameras? Where will all that data be stored?
This example demonstrates why comprehensive knowledge of the technology is required. One issue is that knowledge of the technology leads to the conclusion that raw data from high-definition cameras cannot be transferred online, so protection of the data stored in the car and cryptography of the data to prevent dissemination are issues that need to be addressed before hand.
As you dig in - more issues, such as retention of data, how much is saved and for how long, or issues regarding dissemination of data, come to mind. Even if high definition pictures cannot be sent using WiFi connections, the car can still send the analysis of the pictures, or data interpreted from the pictures, which is more condensed and can be sent on-line, to any data center around the world. How will this impact state security, if information regarding transportation can be transported all around the world?
I therefore have to emphasize in my study, discourse with engineers and software designers. Only through understanding how they solve technological issues, and only after I can understand the architecture of SDVs – I can try and formulate a legal solutions for the new challenges brought up by SDVs. I try to understand why different technological tools are being implemented, and therefore can differentiate what is technologically required and what can be done to alleviate potential legal issues or potential infringements of legal rights.
This is not a simple task. It requires a legal scholar to start speaking “tech”. My material for the legal study is no longer confined to legal articles, legislation or judicial material – but also includes computer blogs, commercial publication of new technologies etc. It requires constant adaptation – trying to make law which is technologically oriented.