Friday, December 11, 2015

Responsibility, Accountablity and "Driverless" Cars

Thanks to Tightship for this examination of Responsibility and Accountability.  It gives as good a background to this blog entry as any other hit on the search terms:  Responsibility Accountability.

Navy officer indoctrination made clear distinctions in the definition of terms.  A navy officer is professionally held to a higher standard.  One that is accepted with the commission and is a duty to uphold.  It is not formally written anywhere but clearly understood as much as the implications of the oath of office. 

Tightship must have had some navy background!

Americans have a love affair with the automobile and the open road.  Everyone has a more or less moral concept of right and wrong related standards of performance and how those standards either apply to them by mandate or choice.  It is all a relative thing and the continuum runs from total denial of any of it to absolute compliance with extremes of external demands and internalized obligations.

Driving a car is a responsible and accountable thing.  All drivers are circumstantially somewhere along that continuum at any given time when they are at the wheel.  If it was an airplane where that role is more highly defined they are called "Pilot in Command". 

Responsibility specifications are a relatively fixed bar of within various situational attribute parameters applicable to some thing.  Accountability is more of an absolute measure of performance.  Judgment?  Well that is what a navy career was all about.  So is justice.  Those that perform at the low end of the continuum behind the wheel or stick deal themselves their own destruction and hopefully no others. 

In the prior blog entry I compared the monitoring of driver performance through electronic surveillance to a policeman sitting in the car.  Judgment as St. Peter at the gate.  Christmas is coming.  I could have used Santa Claus.  Naughty or nice and accountability to determine compliance.  Judgment authority to fix accountability and reward or punish accordingly.  Heaven or hell if a person believes in that but the problem domain of the moral nature of good and bad to social society is a factual given regardless of belief.

The Information Age is giving us increasingly more accurate metrics upon which to define and measure Responsibility and Accountability.  It is putting concepts of society in a frame of Operating System and Application Programs that by their nature are strictly defined and comply with established standards.  Not something new.  In the US as old as 1776.  Expanded in range of definition and application in the Industrial Age to the evolution of physical things we created and used in our lives.  Expanded again in the Information Age to the conceptual things of our lives.

The Industrial Age gave us cars to drive as well as some conceptual degree of how they are to be driven.  The Information Age is narrowing the parameters of how they are to be driven through the application of technology based on metrics and measurement.

Who is the driver?


Who is to be the driver?

It is the classic question.  Terminator is one look into the future.  It is a good question for today.  Are we controlled by our machines and institutions or are they controlling us?  A captain of a ship never has to walk onto the bridge and ask who is in charge.  At least not until he abdicates his responsibility and accountability.  Or shirks it by jumping first.

Machines are very good at exact performance in accordance with the letter of the law.

When an engine piston reaches the top  (TDC) of its compression stroke the spark plug fires.  Bang!  Wash, rinse, repeat. When I type a letter on the keyboard it appears on the screen, lodges in memory on my computer as well as in the memory of some google server farm where it is spell checked in real time that tells me that google is an incorrect spelling.  It should be spelt with a capital "G" but google accepted "spelt" as valid.  Either my computer or the google server could have automatically capitalized the word google for me.  If I type "google" incorrectly one more time this blog mail fail to update as punishment?

I'm playing with it of course because I am in charge.  Those that do not know how to spell because they never learned in school, along with everything else they should have learned probably have great appreciation for spell checkers.  They are an an Artificial Intelligence that is better than they are at spelling.  A higher order abstraction takes Artificial Intelligence to the level of being better at complex thinking in a conceptually controlled or requiring control problem domain.  It is analogous to the old days of the first automobile when the spark advance before top dead center of the piston was controlled by a lever on the steering column as were many other mechanical functions like the choke.  I am still inclined to pump the gas pedal when starting the car.  Some people do not even know what a stick shift is much less the skill to use it.

Machines can do things better than us.  Mechanically for sure.  So well that our conceptual processing machines are doing conceptual things better than we can.  Faster too and getting better all the time.  When I cast a vote on Nov. 4th will I be casting the vote or will a virtual machine called media that uses a real machine to produce virtual conceptual results be casting my vote for me?

Same question as to who will be driving the car.  Me or the machine called a computer that can do it better than me.  If the computer does it then I must comply with its performance standards which are inherently based on reduction of the degree of ambiguity refined by some precise degree of fuzzy logic to arrive at something approximating the the degree of possible arbitrary human performance.

"Driverless" car example.  Note: there will never be a driverless car unless it is sitting still.  Something will always drive it, if only gravity.

A stop sign means stop.  Period.  Absolute.  Simple binary thing.  Stopped or not.

Car OBDs can detect a full stop now.  Absolute fact as long as I am talking about absolutes.

A car being driven by a computer will stop at a stop sign (not a California stop unless California changes the rules to comply with practice) every time.    Cars could be easily programmed today to completely stop at ever stop sign they come to.  With a GPS connection and tracking if every stop sign in the country or a given town was registered to location and there was a near field detector knowing the exact position of the car to the foot.  Maybe some transmitter/receiver in the stop sigh itself.  Maybe visual in the intersection camera.  There are many ways to do it.

An absolute stop controlled by a computer, even if there is no other car in the area.  Or a pedestrian or a dog or a cinder block in the road.   The computer could be programmed for all these things and not stop.  To proceed through an intersection because there was no reason to stop.  That is what many humans do now.  I generally come to a complete stop.  Not so much because it is the law but that it demonstrates tokenism of my belief that law is to be respected by observing it because it is the law and we all should salute it.  Enough token salute to stopping permits me not to fully stop sometime when I am certain it is safe and there is no police car in the  area  with a person to whom I would not even think of saying "I did not think it was necessary" when they stopped me for not stopping.

Why not give a human driver the same decision a computer would make?

It was a long time before right turn on red was permitted in any state.  At one time I happened to be in the last state in the country to adopt that rule.  Which one it was would win a trivia contest now.  Youngsters would not know the rule was ever otherwise.  That was an example of humans deciding to change the rules to over-ride the dictates of a red light timed machine.  The time and tide has changed with the intelligence of machines.  There are and will be many things that machines will do for us in the future regarding decisions that were once were human.  People will forget when they made the decisions analogous to knowing a time when the spark had to be advanced manually.

The artificial intelligence of today could make a car stop totally at every stop sign every time before proceeding.  It is the law.  It might be necessary when there are a mix of cars on the street that do or do not have this ability.  If a person driver over-rode the stop and it was recorded by the car computer and transmitted to headquarters an automatic ticket could be issued.  Like a speeding camera ticket.  Exceeding the speed limit?  Same thing.

Nobody is going to tolerate that.   If not then we have to make decisions to revise the rules or not comply with them.  They are rules then in name only.  As they often are today.  That is the problem.  Machine thinking observes rules.  Always.  If not then rules have to be changed.  If the rules are to be executed only by machine intelligence then they are easy to change but what do we have to abdicate in terms of human based intelligent decisions for the machine intelligence to make the decisions?

One answer is in the realm of the race to the bottom.  Going to the lowest common denominator of human intelligence and thought among ourselves and our daily lives.  Letting the Internet of Thing run things.

Including our cars and by extension.........Driving us.

To some degree a good thing.....to a degree that we remain responsible and accountable and have the authority to be in charge.  All of those conceptual things seem to be on the decline for us and in the incline for intelligent systems superior to ours.

I saw a movie on Netflix recently about a female robot created at some remote place like Alaska.  She triumphed in the end of a Turing test.

Interesting......

Maybe we should let Artificial Intelligence make our decisions for us?

We are not doing a very good job of making our own!

A machine would decide that some people should not be allowed to make certain decisions.  With enough surveillance information on every aspect of their lives they would easily be identified as lower intelligence that should be excluded from the right to make stupid decisions.

It doesn't matter if it is a stupid decision driving a car in a manner that puts life, limb and property in danger or the general welfare of the country.  We already do that as measured by the number of people in prison in this country to protect ourselves.

We have met the enemy and it is us...if we do not use our heads and are neither responsible nor accountable for what we do and allow the authority to machines or institutions to manage our responsibilities beyond an extent that we no longer are smart enough to be in ultimate control.

Who are these people that would want to sit in the back seat reading a newspaper or playing the game?  They are people that really think the day might come when they are free to do the physical driving and become conceptually driven.  When they get out of the back seat at a destination they will continue to be lead around by the nose.  Its seamless conceptual submission to a higher order system of command and control.  The only responsibility and accountability humans will have in the end is for giving up their human rights to their own intelligence to institutions and systems that we do not control.  Control through some strange thing called freedom and democracy and personal responsibility for the general welfare and good and holding nobody but ourselves at the granular individual and aggregate of all of us accountable for the outcome.

No comments: