4
$\begingroup$

I'm doing an EPQ (Extended-Project Qualification) on Artificial Intelligence Bias, and would like to gather some primary data for analysis.

Do driver-less cars lead into controversies that are to do with ethics? For example: if the AI had to choose to minimize damage and could not avoid casualties.

$\endgroup$
  • 1
    $\begingroup$ I can see you edited the old question, then copied it to a new one. That is not really necessary - the old question gets reviewed following your edit, and you would get feedback there that might help refine it further. However, now you have done this, I think it is better to move forward with this copy. $\endgroup$ – Neil Slater Nov 26 '18 at 19:40
  • 1
    $\begingroup$ @NeilSlater There do seem to be ethical issues fairly unique to autonomous cars. (Relates to this prior question: How would AI prioritize situational ethics? but I think the narrowing of scope makes this current question new.) $\endgroup$ – DukeZhou Nov 26 '18 at 19:42
  • $\begingroup$ I'm reading through the website of the University of Manchester which explains what an Extended-Project Qualification is. Because I belief, that the asker has included this information in his question with some reason. The question is, what was the reason? My assumption is, that the course is not about Artificial Intelligence, but about teaching Artificial Intelligence to the public. So it's a course about journalism or how to become a teacher, right? $\endgroup$ – Manuel Rodriguez Nov 26 '18 at 22:52
  • $\begingroup$ Sorry, I'm new to the site; and regards to the reason, it is a topic that you choose yourself to write a report of 5000 words on. It focuses on the skill of research and analysis, as well as concluding your findings and making sure there is no bias. I asked the question due to gather some primary research but my first question was rejected as it opened debate. I'm not entirely sure how to phrase it well, so it's a bit hard for me. $\endgroup$ – Jia Xuan Ng Nov 29 '18 at 20:42
1
$\begingroup$

Ethics of Research Into AVs

Research that does not, in the course of lab operations, harm people in any stage of development or mammals used in lieu of people, rarely present prohibitive ethical conflicts. Whether a perception of possible ethical conflict limits funding and presents marketing and public relations challenges to those who sell the products of engineering using the research output is a different question.

Productization Challenges

Initial safety metrics and the media play of accidents that must inevitably occur and how these indicators compare with human driving metrics will have much to do with early perceptions. The automotive industry and new entries into it are aware of this. That automotive products are such a ubiquitous part of modern life both raises the stakes in terms of profitability and will increase the visibility of the transition.

Driving Ethics Defined by Law

Statutes and case law have already clarified the ethical expectations of the public regarding safety priorities. Early public discussions about public safety and the introduction of automated vehicle AI shows no significant conflict between the straightforward application of existing law and public expectations of full driving automation safety. Saving lives comes before saving people from non-fatal injury, which comes before saving pets and property. The trailing priorities are saving time, saving fuel, and saving wear on the vehicle itself.

A search on this AI site using the tags associated with AVs, self-driving, and ethics will reveal much more on this topic that is unnecessary to duplicate here. Also a basic search for the word VEHICLE reveals a number of good postings. As an example, this question could use a few answers, but the question in some ways addresses this question: Recognition and Response Generalizations for Autonomous Vehicles or Not?. This question has some of the best discussion on priority in situational ethics in AVs: How would AI prioritize situational ethics?.

The key to analyzing the ethics of the driving process itself is statistical comparison. AVs will not likely be accepted well by the public or by government unless the accident rates reflect AV safety advantages.

One metric that should be primary is the percentage of fatalities among all reported accidents as a function of what percentage of the drivers were computers. Whether legislators and regulators will pick up on the importance of this specific metric remains to be seen.

If statistically proven safer in terms of fatality and injury metrics, then it would be unethical to let people continue to drive cars. Who would dare oppose measures that save human lives, including pedestrians, children walking to school, and the elderly and disabled? This nearly inescapable reality is the foreshadow of a likely future legal paradigm shift.

Normal Resistance to Change

The word I hear most often about fully automated vehicles is the word, "Creepy." Sometimes I must explain to those not up on the AV technology that a fully automated vehicle is like a quiet limo driver, responding appropriately to the specification of a destination and commands to stop and go, possibly asking for clarification on occasion. That they then respond with, "That's creepy," is not surprising. The modern world is increasingly emotionally and economically entrenched in daily driving.

It is important, when considering market acceptance, to notice that the word, "Creepy," is also used for people who are considered potential serial killers, rapists, abductors, vampires, and other fictitious or real types that stalk and either kill or otherwise destroy the psyche of people. Interestingly, being spooked by the suggestion of cars without human drivers is the same response, adjusting for changes in common linguistic expression, as upon the initial introductions of a long list of other technological advancements.

  • From scribes to printing presses
  • From staying clear of the deep ocean to circumnavigation
  • From kerosene to electric lighting
  • From bird watching to human flight
  • From sail power to steam power
  • From horses to cars
  • From tellers to ATMs

Automated home product store cashiers and automated check depositing at banks are still in the midst of a slow and reserved acceptance curve. Yet there are more articles being written and uploaded each day about online social network addiction than there are about heroin addiction. People graduate with degrees in English Lit and find ways to get onto software engineering teams so they can afford a house, and intercontinental sex involving three of the five senses is on the rise. In general, postmodern people in most economic and education levels worldwide are almost risk-oblivious toward most technology changes.

Looking Back from a Possible Future

Another ethical question may appear in hindsight in another century. Those future people may read about the driving conditions of today and wonder how nineteenth century people decided to ignore the ethical considerations of allowing unqualified and often highly distracted people to drive heavy equipment. (That is a culture-bias-free description of what much current car and truck driving is.)

When the transition from horses to cars began, cars were relatively slow. Still, people saw that cars could run over children and the elderly whereas a healthy horse would not. The genes of the horse that would trample a child had been eliminated by putting down horses that trampled children over a period of thousands of years. So in historical context, it may later become an ethical problem to continue to let humans drive cars.

Ethics in All Technology Introductions

The introduction in technology generally creates jobs and eliminates others. What the net effect is depends on the technology. Brewing beer created many jobs and displaced a very few wine makers, producing a net improvement in employment opportunity. Farm automation displaced many small farmers and created very few equipment operation and repair jobs in comparison.

The displacement of professional drivers may be an ethical problem. Whether some private car owner lets their car drive them to the store or yoga class is not challenging any ethical standards, except for the ethical perspectives of Ludites, Amish, technophobes, and other fringe groups. However, buses and trucks lack some of the frequency of challenging choices that cars, trucks, recreational vehicles, and minivans have. The probability distributions for lane transitions, cruising speed, acceleration, stopping deceleration, stopping conditions, turns, route deviations, and signalling are narrower. They are much more likely to be fully automated first, and those drivers have no lateral moves within those transportation companies.

Since both licenses and registrations would be held by the same entity, the AV itself, motor vehicle department workers will be displaced.

Although not entirely displaced, some businesses and careers would likely be negatively impacted.

  • Personal injury lawyers and legal teams
  • Insurance carriers, agents, and administrative organizations that handle a large number of automotive policies and claims

Ethical Considerations Other Than Employment

A number of systems that are based on universal social standards related to cars will likely resist change to computer driving and may launch campaigns opposing the public relations campaigns launched by those introducing AV products. Such opposition would likely use mechanisms to build apprehension about possible collapses of social infrastructure. There are several groups that benefit from human driving. Most of them are the personnel and systems that use driver's licenses as a primary tool to support law enforcement and investigative work.

Acceptance and Public Relations

Convincing people may be about clever marketing.

  • If you saw all the crazy driving in my city, you wouldn't be quibbling over removing steering wheels.
  • Keep teens safe.
  • What's the real objective? Driving or getting there?
  • AVs save fuel.
  • Even fewer accidents with the Honda Driver-free Autobot
  • Like having a limousine service in your garage

One of the greatest feats of marketing and public relations is the construction of the idea of personal transportation. Betting that automotive manufacturers can overcome the initial sense of creepiness and sell the new AI infused products is not exactly a bet for fools. Consider their current achievements.

  • A large percentage of families in nearly every income bracket with one personal vehicle per adult family member
  • People spending 20% of their annual income solely to own the latest trending body aesthetics and dashboard features
  • Commuting through hordes of traffic to get to a cubicle containing a computer no more powerful than their home computer and sitting on the same Internet
  • People selecting their life long partners on the basis of the car or truck that person owns
  • The assembly of a sustainable and unconscious social force to buy a new car every two or three years when fully restoring the current one would require less than a quarter the expense
  • People, based on an association between mechanical power and social dominance, maximally accelerating to just over the speed limit merely to stop at the next light in an only slightly different order, modifying their arrival time by only two or three seconds in exchange for unnecessary mechanical wear and burned fuel
  • The creation of a global economy as dependent upon cars as the availability of air or water

The automotive industry even surpasses the fashion industry in their public relations and marketing prowess.


Footnotes

Other Aspects to This Answer's Approach

No formal definitions of the terms used in the question were given.

  • Extended-Project Qualification
  • Artificial Intelligence Bias
  • Primary data for analysis
  • Controversies that are to do with ethics

It was assumed that gauging whether computer driven vehicles will be a technology that achieves ready market acceptance or stalls due to apprehension is the objective of the question author. Until more specifics are given about what would make data primary and to what decisions the qualifications mentioned would apply, hunting for data sets would be premature.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.