The problem I have with this analogy – between self-driving cars’ behavior and decision-making capability (or human drivers’ for that matter) – and the trolley problem (https://en.wikipedia.org/wiki/Trolley_problem), is that the trolley problem presumes a “runaway trolley” – i.e., one that is not in control.
Cars, whether driven by humans or artificial intelligence, should not be going too fast for conditions in the first place, and, if they are not, then there should not be a decision that has to be made between the safety of the occupants and the safety of other road users outside the car. Speed limits are a maximum legal limit, not a minimum, and there are any number of conditions that warrant going less than that maximum legal limit. But until we stop designing roads to feel safe when driven at twice the speed limit that will be posted (or more), this will continue to be a widespread problem. The only time I ever see all the cars on University Ave through campus going 25mph or less is when congestion there severely limits the free flow of motor vehicles. Yet removing the congestion there (the only condition that actually seems to decrease vehicle speeds to what is considered legal and safe) and restoring free flow of motor vehicles is approached as a problem to be solved. Chuck Strawser Pedestrian & Bicycle Transportation Planner Commuter Solutions UW-Madison Transportation Services Please note that my email address has changed. My new email is [email protected]<mailto:[email protected]> Visit our University Bicycle Resource Center at Helen C White: http://transportation.wisc.edu/transportation/bike_annex.aspx How are we doing? Take our customer satisfaction survey at https://www.surveymonkey.com/s/CommSol_CSSurvey From: Bikies [mailto:[email protected]] On Behalf Of Hank Weiss via Bikies Sent: Sunday, October 30, 2016 11:45 AM To: 'Bikies' Subject: Re: [Bikies] Mercedes-Benz has a algorithmic solution to the "trolley problem" in the age of self-driving cars Scott: Just to muddy the waters a little, it is noted the CEO of Mercedes clarified his comments later: Fortune later reported, that Mercedes-Benz said that their CEO, von Hugo, was misquoted. They later sidestepped the question saying the automaker’s official position is that “neither programmers nor automated systems are entitled to weigh the value of human lives,” and that the company is not legally allowed to favor one life over another in Germany and other nations. Interestingly, an earlier study by Bonnefon et al. in Science<http://science.sciencemag.org/content/352/6293/1514> magazine reported that the majority of the people surveyed thought it would be ethically better for autonomous cars to put their occupants at risk rather than crash into pedestrians. I would imagine this to be the view of most pedestrians, cyclists and among ethicists who would point out that the hazard itself is created by the large heavy vehicle traveling at speed (with better occupant protection) and so should bare responsibility. Yet the majority also said they wouldn’t buy autonomous cars if the car prioritized pedestrian safety over their own so it’s a good indication how the industry might want to handle the issue given the freedom to do so. Others have pointed out that the variants to the so-called Trolley Problem<https://en.wikipedia.org/wiki/Trolley_problem> will become less vital if overall relative safety advantages to other road users from autonomous cars are clearly established despite our ethical and legal discomfort with a small fraction of unusual circumstances. I would point out fully autonomous systems are still quite a ways off. They cannot yet drive in snow (though micromapping and positioning makes that theoretically possible), have trouble with detours and unmapped changes, can’t tell a large pothole from a puddle, a dog from a child, and can have problems with shadows and rain. And whether they will find success in the market remains to be seen. People may be emotionally reluctant to hand over control of their vehicles when transport interactions are algorithmically driven. Case in point, assuming cars will be programmed according to each states laws (a nightmare in and of itself) how will motorists feel when their car robots actually do stop at all crossings where a cyclist or pedestrian is waiting to cross (as opposed to current widespread obliviousness to this obligation)? And what about actions where people walk or bike or put objects in front of autonomous traffic just to trigger the avoidance strategies for fun? Personally I think different parts of the smart driving universe will come in piece-meal (like vehicle-to-vehicle communication), or in very specific places like freeways long before we see them allowed without driver oversight on local streets. Cheers, Hank Weiss From: Bikies [mailto:[email protected]] On Behalf Of Scott Morris Rose via Bikies Sent: Saturday, October 29, 2016 10:10 PM To: Bikies <[email protected]<mailto:[email protected]>> Subject: [Bikies] Mercedes-Benz has a algorithmic solution to the "trolley problem" in the age of self-driving cars In a variant of "the customer is always right" principle, they code to - surprise, surprise, surprise! - save the occupants of the vehicle in favour of any number of pedestrians or other road users. http://blog.caranddriver.com/self-driving-mercedes-will-prioritize-occupant-safety-over-pedestrians/ -- S. Rose
_______________________________________________ Bikies mailing list [email protected] http://lists.danenet.org/listinfo.cgi/bikies-danenet.org
