Mike__S

"Lawyers will always attempt to assign individual blame. Personally I don't accept the notion of large scale collective error. Just as with the General Motors ignition switch fiasco, defect was traced to an engineering design team and the QC engineer who approved that design."

This General Motors example would seem to be, if not "large scale" (your words), a collective error, if ever I heard of one. A process co-dependent on the cultural environment. But, it is also beside the point. As I reflect, the point is not really about assigning responsibility for error, but a question of a systemimic responsibility for individual outcomes and I have been incorrect in specifically using the word "error".

This situation ought to be better thought of more as a collective recognition of relative risk acceptance. And as opposing a lawyer arguing for absolute risk avoidance because of individual negative outcomes. This is much more a question of policy than common law, and it would seem prudent to indemnify large corporations that provide workable solutions. It is early times yet, so this thought may seem inappropriate to some minds, yet just such a policy decision was made announced today by NHTSA in the Florida Tesla fatality.

Such systems (drive assist) must be understood as to have a collective or random risk factor, for which individual specific negative outcomes cannot be assigned a direct cause and effect relationship. The collective benefits vastly outweigh with the individual risk, i.e. we are acknowledging that a huge number of accidents and injuries that occur today will not occur when these systems are in wide spread use. In this sense we must learn to accept the collective and random risk factor inherent with using the system.

I think the automated safety, and by extension, autonomous driving (not the other way around, BTW) is far more subtle than ferreting out individual responsibility, and that the bug splat analogy is most appropriate, in fact critical, in understanding the import of this new technology. The question is not of error or assigning responsibility, so much as group agreement that certain (new) types of risk are worth taking.

My prediction is personal injury case law will eventually accommodate the bug splat principle, an the aggressive seeking of "individual blame for error" responsibility will be augmented by the results this collective risk and the vast numbers of positive outcomes active safety achieve, despite inattentive drivers making serious errors. Drivers do enough of that already, don't you see? The aim is to reduce the huge numbers of negative outcomes of that risk population at very negligible risk to the general population of drivers.

Mike__S

"We ought to keep this reason for active safety developments foremost in our mind's eye. It is the variation in driver attentiveness and downright quality of judgement that is the core issue. Some of us, at least some of the time, need all the help they can get."

"Unfortunately, the industry is creating new technology for "convenience" operations rather than for safety. Systems are available to detect and prevent impaired drivers from operating a car, but they are not a feature in any car that I know of."

Here, we simply have corporate lawyers doing their best to avoid hot water. This is best understood by the comments I made in my post just previous to this one. This is a particularly sensitive issue in the United States, the most litigious nation on earth. Other markets do not have quite this sensitivity, for broad cultural and legal reasons.

Of course these technologies are for safety, but because of the gray area of responsibility, it must be addressed with double speak at present. Maybe George Orwell had it right. Everyone knows the expression: "What kind of stupid do you think I am?!" For those individuals that do not see through this connivance, I have a bridge in Brooklyn I like to sell them, too.

Sweden and Finland (maybe Norway and Denmark?) all have vehicle starting devices that test for driver alcohol content. This is because, like Russia still today, they once had a huge drunk driving problem. Volvo cars I know can be ordered with pre-wiring, just as was once done for cell phones and audio equipment.

Drunk driver convictions in Scandinavia (and elsewhere) now have huge penalties. Really huge, commensurate with the economic status of the individual and a requirement that any car licensed to that person's household subsequently be fitted with this device. As a result, even many bars in Scandinavia (and UK last I was there) now offer a breathalyzer test for departing patrons. Once attended a three day New Year's party in Sweden decades ago. Only time I have managed to stay wasted for 72 hours straight. It all makes sense to me.

The issues the the USA that preclude rapid adoption of such devices are the fact that each state has jurisdiction over driver licensing and the fact that except for Louisiana, they all use Common Law. The Common Law is basically proscriptive, where anything not forbidden, is not, defacto, illegal. It is illegal to drive your car while drunk, but it is not illegal to get into the drivers seat and start your car while drunk, as this in itself cannot be proven to cause third party injury. This is how the argument could go in Common Law courts.

It involves a lot of time and money to get past such case law related "liberty" arguments. This is changing over time as case law does change, but in principle this is the why many seemingly clear policy issues are hard to put forward quickly and cleanly in Common Law. Pollution cases were famous (at one time) for the maneuvering that went on.

Roman Law as used in Europe (outside of UK) is basically prescriptive, where anything not permitted can be administratively judged illegal, such as getting in your car and starting it while drunk. No trial, no plaintiff/defendant, no lawyers, at least as they are known in the States. So, with nation wide driving law, and a prescriptive Roman Law tradition it is quite simple to enact such seemingly draconian laws in Europe. And, they work, too!

If you extend this example to active safety, that is this Common Law vs. Roman Law difference, you can see exactly why the auto makers are treading on thin legal ice in the US. Whereas in Europe, the extent of the liability is broadly already established. Thus, offering obvious safety technology advancements are met with, shall we say, a more rational and measured response on the part of government and the public. Different system there and different attitudes totally.

On the other hand, I do not want to appear to be glossing over individual problem experiences reported here. These systems are new, some real improvements have yet to be discovered, and some serious incidents will undoubtedly still occur. What I try to do is stand back and illustrate the broader and longer view of what is implied by implementing this technology.

In a sense, it is a very brave thing early adopters are working with this technology. I will soon be doing so with the confidence that it is a worthwhile risk for me to take. The chances of this technology preventing accidents by me are overall of much greater value to me as an individual than the risks implied. The hope is that eventually, with the majority of cars fitted with active safety, we will all be much better off.

Mike__S

The nature of your question is foremost a question of semantics. To call this phenomenon a software bug is a misnomer because of a misunderstanding of how these systems function.

The active RADAR or LIDAR would still see a person, animal or broken vehicle and alert you accordingly, but the vision camera that detects guidance and signage recognition would not react to shadows of a building or overpass at a given time of day on each day of the year and thus give a false positive. Geotagging is precisely this accurate.

This is a good example of how sophisticated and necessary (externally recieved) meta-data tagging is and why it is so important for even solid level two automated guidance systems, much less moving forward to semi-autonomous level three or fully autonomous level four.

Meta-data tagging is quite advanced in places like Germany and elsewhere in portions of Europe. For the US, in districts heavily traveled by Tesla vehicles, there is also some good meta-tagging. In Germany, the meta data is widely available to all manufacture's systems, while in the US this process is only in the early stages of deployment.

Hope these details help everyone.