Robots and Wisdom

It may well be that the time has come to make three fictional laws, actual laws in our world. Perhaps the United Nations could be given authority over Artificial Intelligence and could enforce Isaac Asimov’s “Three Laws of Robotics.” Two articles in the news today bring to mind the need for greater regulation in the areas of Artificial Intelligence (AI) and “outer space frontier development."[1] We now need the “laws of robotics.” You may not remember these famous literary laws from Asimov’s short-story, “Runaround.” Allow me to refresh our memories:

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.
2. A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.
3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.[2]

These laws would prevent the kind of military drones now being contemplated by the Pentagon as they discuss policy with Google and other leading AI industrialists.[3] It would be very easy for drones and AI drones to take over conventional warfare. Such laws, if written into international law, would have the potential to demilitarize the AI and Space frontiers and prevent us from reverting to a “wild-west” mentality in these spheres of influence. The UN (including the little known United Nations Office for Outer Space Affairs [UNOOSA]) would need to be given authority to oversee and gain cooperation from member nations - and therein lies the difficulty. The recent launch of a swarm of private artificial intelligence satellites, each about 10 centimetres across, demonstrates how difficult it is to enforce the laws. These recently launched satellites are so tiny that they cannot be tracked and could pose a navigational hazard for the International Space Station (ISS) and other satellites encircling the globe. Swarm Technologies was denied permission to launch these satellites from within the United States, so they found a carrier in India willing to send them aloft.

So, if we can’t enforce the rules, laws, and guidelines already in place, how could we enforce the three laws of robotics? Yet, it seems we must try. My recent Bible readings addressed the topic of “good judgement.” What does good judgement look like in some of these recent developments? I leave that for the reader to meditate upon as we read the following excerpt from the book of Proverbs.

Proverbs 4:4-9

⁴ My father taught me,
“Take my words to heart.
    Follow my commands, and you will live.
⁵ Get wisdom; develop good judgment.
    Don’t forget my words or turn away from them.
⁶ Don’t turn your back on wisdom, for she will protect you.
    Love her, and she will guard you.
⁷ Getting wisdom is the wisest thing you can do!
    And whatever else you do, develop good judgment.
⁸ If you prize wisdom, she will make you great.
    Embrace her, and she will honor you.
⁹ She will place a lovely wreath on your head;
    she will present you with a beautiful crown.”

---

[1]See, “How we can prevent outer space from becoming the Wild West?”; Nicole Mortillaro, CBC News, Jun 17, 2018, http://www.cbc.ca/news/technology/space-regulation-1.4687587

Also see, “Is Google's new set of principles enough to calm fears over militarized A.I.?”; Ramona Pringle, CBC News,  Jun 16, 2018, http://www.cbc.ca/news/technology/google-militarized-ai-1.4707697

[2]Asimov, Isaac (1950). "Runaround". I, Robot (hardcover) (The Isaac Asimov Collection ed.). New York City: Doubleday. p. 40. ISBN 0-385-42304-7. This is an exact transcription of the laws. They also appear in the front of the book, and in both places there is no "to" in the 2nd law.
[3] See, “Is Google's new set of principles enough to calm fears over militarized A.I.?”; Ramona Pringle, CBC News,  Jun 16, 2018, http://www.cbc.ca/news/technology/google-militarized-ai-1.4707697

Pedestrian Fatality and Autonomous Cars

(Click on thumbnails for larger image)

A recent pedestrian fatality involving an autonomous car is simply the first such tragedy. Science News reported that a woman died crossing the street when a self-driving car, deployed by Uber, hit her.[1] Writers, engineers, and computer science experts had predicted this for some time. What will this mean specifically for the autonomous car industry and artificial intelligence in general? What algorithms, if any, need to be rewritten? What logic did the self-driving car’s onboard computers use in choosing to enter the area when a human being was also in the street?

I have previously discussed this possibility and its implications in other posts.[2] A number of questions now come to mind. Did the car make a decisive choice to hit the pedestrian to avoid harm to those in the car? Could the onboard AI have made a different choice? How could the outcome have been changed? Investigators will need to ask these and several more questions to get to a place of assigning responsibility for this accident. What will the law have to say? Who could be fined? What sort of lawsuit could be filed? We await the outcome to see how manufacturers of autonomous cars might learn from this incident and the ramifications for all car makers.

---

[1] http://bit.ly/2FOICfm

[2] http://bit.ly/28VmLE1 and http://bit.ly/2AsuDZI

Autonomous Cars

Self-driving cars, without humans in the driver’s seat, are being tested on the streets of some cities. I used to joke with my nieces that once they were driving I would be turning in my driver’s license. Now, I am not sure whether I should be filled with awe at this amazing autonomous technology or fearing for my life. There are still many questions about how robots will make critical choices (see Robotic Laws). The tech companies readily admit that putting autonomous cars on the streets is a form of beta-testing and that there will be accidents involving cars without human drivers. However, the rationale for putting autonomous vehicles on the streets is that, ultimately, robot cars will lead to fewer accidents and fewer traffic fatalities on our roads. This is the argument in an article in Science News: “When it comes to self-driving cars, what’s safe enough?” by Maria Temming, November 21, 2017.

What kind of backlash against these autonomous cars can we expect when an autonomous car collides with a car piloted by a private citizen or professional driver? Temming gives us a possible scenario and asks, “What happens when a 4-year-old in the back of a car that’s operated by her mother gets killed by an autonomous car?” This is a real possibility. Will the public be quick to blame the robot or the human? How much will driving conditions factor into the investigation? How will Canadian winters affect autonomous cars? The title of the article is indeed the pertinent question, “What’s safe enough?”

Robotic Laws

Isaac Asimov's "Three Laws of Robotics"

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.
2. A robot must obey orders given it by human beings except where such orders would conflict with the First Law.
3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.[1]

Conversations about driverless cars just got a little more difficult. Isaac Asimov's "Three Laws of Robotics" cannot help robotic cars in situations where they must choose between harm to passengers or harm to pedestrians. A recent article[2] in Science News includes the following graphic and explanation to illustrate the point.

J.-F.Bonnefon et al/Science 2016

"Driverless cars will occasionally face emergency situations. The car may have to determine whether to swerve into one passerby to avoid several pedestrians (left), swerve away from a pedestrian while harming its own passenger (middle), or swerve away from several pedestrians while harming its own passenger (right). Online surveys indicate most people want driverless cars that save passengers at all costs, even if passenger-sacrificing vehicles save more lives."

Of course one might argue that, on average, humans are no better at such moral dilemmas. Perhaps the issue is that with a robotic car we know exactly what we will get. Programming will determine the choices a driverless car will make in any given situation, whereas humans are much less predictable and the choices they make are dependent upon the degree of altruism resident in the individual driver. Many times, in a moral dilemma, we do not know how an individual human will respond. Will they choose to sacrifice themselves for the sake of strangers? Will they sacrifice themselves and their child for the strangers? What about the scenario in which their pet would be sacrificed for the sake of unknown pedestrians? We may do surveys and get a statistically accurate average on the answers to these questions but we know that there would be outliers and unique decisions made at the spur of the moment. With robotic programming, the driverless car does not make a choice, it simply follows the program with which it was built. This puts the onus on the manufacturer rather than a driver or a car or a robotic brain. In a litigious society, this may be the problem that slows the progress of driverless cars.

---

[1] In 1942, the science fiction author Isaac Asimov published a short story called "Runaround" in which he introduced three laws that governed the behaviour of robots. https://en.wikipedia.org/wiki/Runaround_(story)

[2] “Moral Dilemma Could Put Brakes on Driverless Cars,” Bruce Bower, June 23, 2016, Science News, https://www.sciencenews.org/article/moral-dilemma-could-put-brakes-driverless-cars