Reflections on FOSS and the Automotive Future

-- By EmiL - 07 Nov 2016

Jerimiah Foster’s presentation, and the subsequent discussion, on open software in autonomous cars raised three strands of thought related to Free Software licensing, individual rights, and the intersection of the two. First, rising expectations for content delivery to our cars has led to complications in compliance with GPL. Companies that are using software licensed under GPLv2 may be able to circumvent the spirit of Free Software with restrictive hardware. This desire to prevent others from being able to modify their software leads to a second problem of regulation and repair. As more Free Software is licensed under GPLv3 companies may move towards the development of proprietary software to avoid the anti-tivoization clause. The unfortunate result of a rise in proprietary software would be a decline in transparency. The use of proprietary software in cars raises serious questions about the ability of regulators and citizens ability to feel safe driving around a thing that knows, hears, and remembers everything. This leads to the third matter that was raised during Jerimiah’s presentation— the rights drivers. I will spend the remainder of the essay discussing my initial thoughts and open questions on drivers’ rights and what the future will likely bring in reality.

Here, There, and Inbetween

All drivers must have the basic right to start and stop. As self-driving cars becomes more acceptable, there is a real danger that a car could take you or keep you at a location where you do not want to be. Thus, a right to start and stop would allow a driver to shutdown the software entirely and manually drive the car to one’s intended location. The absoluteness of this right is complicated in two ways. First, the reality is that the software that controls the driving mechanisms also runs the required safety features of the car. Much of the appeal of autonomous cars is the capacity to create safer driving conditions, so I would foresee strong pushback against a manual shutdown feature if it came at the price of disabling other, ostensibly necessary, safety features. Therefore, keeping human autonomy as the top priority may help to shape the development of car software so that the shutdown of automated functions doesn’t disable desired features. A second complication is the disparate legislative progress across states when it comes to the regulation of autonomous vehicles. For example, California passed a bill this summer that establishes a pilot program for cars that have no steering wheel, accelerator, or brake. If people accept pilot programs like California’s as positive progress towards safe roads, by the time Congress decides to take up the issue of autonomous vehicles again, the tide in favor of hands-free cars may be too strong to require that all cars have capacity for complete human override. As far as rights go, the right to not be held captive by your own car seems obvious. But it also led me to think about what happens in between points A and B. I believe just as important as getting to where you want to go is that your route is not dictated by third-party interests—a right “of way.” Even when your car is taking you from point A to point B as directed, you as the driver ought to retain full control over the route from A to B and not be subject to the path that has the sponsored car wash on the way. Some may compare this sort of directed-driving to targeted ads on the net, but there is something far more invasive and coercive about beging dragged by the health food store under the guise of going to the gym. In reality, it seems like we are headed in that direction and a right of way will be understood as privilege. Given this reality, the question then becomes whether opting out of ad-based directions will be free and simply require jumping through a few more hoops or if it will look more like paying for a commercial-free Hulu account.

Privacy and Ownership through the Lens of Safety

A discussion of the level of control drivers have over their route naturally leads to the questions about rights as they relate to the data supporting the system. Even through the relatively uncontroversial frame of safety, it doesn’t take much to see that autonomous cars raise new potential conflicts over ownership and privacy. Even those most open to the universality of autonomous cars will inevitably ask, what data belongs to me? If the driving force behind autonomous vehicles is that they will make roads safer for everyone, then it seems to logically follow that data collected about things like if I ever speed in school zones, whether I am cautious near construction zones, and at what times of day I tend to drift into other lanes, should all be transferable to my next car. I believe safety-related data a car collects ought to be thought of like medical records. Every person, even if they choose to drive a different make of car is entitled to have their care operate as safely as possible according to their individual habits. Manufacturers may charge a reasonable price for the processing the transfer of this information, but that should be the extent of costs related to safety data.

This perspective, however, becomes problematic for two reasons. First because it suggests that ownership is dependent on the nature of the data rather than the existence of an independent right for a driver to control his behavior data by virtue of the fact that he generated it. There is no way to draw a clear line between “critical” information that warrants direct consumer control and what secondary information manufacturers can keep for themselves. Ideally, there would be no need for a line and everyone would have complete control over device-generated data, but if smart phones are any indication of how autonomous cars will be treated, that simply won’t be the case. Extending the ‘safety-first’ approach beyond issues of ownership, it also opens the door to justifying extreme government intervention. Under the banner of safety first, you could imagine a world where the police could override driver controls in a high-speed chase or a routine traffic stop. Or perhaps law enforcement could pay to buy-in like other advertisers to have your car take you on certain routes to make surveillance a little easier. In short, using safety as the ultimate end-goal of autonomous cars will likely lead mean the end of autonomous drivers and as new technology develops, regulation must continue to be anchored in principles of individual rather than collective rights.

Wouldn't we need to begin by asking what a "driver" is? When you use other modes of public transport, you don't expect autonomy. The idea that a "car" is a form of transport that involves exercising rights, unlike a bus or a train and more like a bicycle or a skateboard, isn't somehow inscribed in the order of the universe. If transportation is a service, how does modality change the rights of the service's recipient?

I think you want to work further away from Jeremiah, and further away from GPL as well. The questions you want to ask, and which the next draft can more tightly focus on, have to do with much more general propositions. How should transport be organized in a networked society so as to preserve rights that people have, whether they are "drivers" or "passengers"?