I’m not worried that human beings working on artificial intelligence are going to produce machines that are smarter than us, and certainly not smarter than my three grandchildren, ages 15, 9 and 5, all of whom are definitely smarter than I. I’m afraid that human beings designing artificial intelligence are, in doing so, going to make most other human beings using the artificial intelligence more stupid.
The logic of the human experience won’t be reproduced in the machine as it is for example in railroad interlockings. Rather, human intelligence will conform to and be constrained by the limits of the algorithms with which it engages daily.
Example? Did you hear the one about the locomotive engineer who blew by the stop signal but blamed his speed control system for not sounding the overspeed alarm when he hit 18 mph on the restricting cab signal, when everyone knows you are not to exceed 15 mph while operating on a restricting? I am not making this up.
Is this a real fear? Are these the complaints of an old man feeling bypassed and excluded from the marvelous changes taking place daily? Both?
I’m sticking with “both.” True story time: Once upon a March 2007 if I recall correctly, I attended a conference of the U.K.’s Adhesion Working Group in York, England. (York, provincial capital of the Roman Empire, occupied by the Vikings in the 9th century when Ivar the Boneless—the Viking version of the Spinners’ The Rubberband Man—attacked on All Saints Day and caught the leaders of York praying. York, subdued again but not as easily by the Normans in the 10th century, built an important railway center in the mid-19th century. George Hudson, a British railway financier and politician who, because he controlled a significant part of Britain’s railway network in the 1840s, became known as “The Railway King,” built it.)
And in the 21st? There is a high-speed service (MAS about 200 kph) between York and London, where another meeting was scheduled. A colleague, then working at a firm spun off during the privatization of British Rail, accompanied us and set us up for the high-speed train departing 0800 hours, via LNER (London North Eastern Railway), arriving London Kings Cross at 1037 hours (I think).
The colleague had explained to me how advanced the U.K. system was, with its computer controlled dispatch that automatically made decisions regarding routing and what we used to call the superiority of trains, but is now known as “priority.”
“Including the terminals?” I asked.
“Including the terminals,” he replied. “The dispatcher only intervenes if there’s a serious disruption or breakdown.”
I didn’t say anything. I didn’t like it. I’d been told for years that sooner rather than later computers would replace me. As a matter of fact, the person who first told me that was a consultant Metro-North’s VP-Ops had hired to plan the design and consolidation of the train control and traffic management functions of the entire railroad into a single location in Grand Central Terminal.
“Who’s going to be running the terminal?” I asked.
“The computer will,” he replied.
“And when we have to change track assignments, alter crew runs and reschedule equipment turns, who’s going to do that?” I continued.
“You won’t be doing that anymore. That will be done upstairs in the control center, by the chief,” he said.
“Never happen,” I said.
“No face-to-face. You have to tell crews, the mechanical department, everybody, directly, face-to-face what you want done. Otherwise it’s just another phone call to be ignored. Besides, dispatchers don’t want to deal with crew runs, passenger counts, equipment cycles, labor agreements, terminal configuration—boring stuff like that.”
“They’ll have to learn.”
He was from Britain, too. I was a Chicagoan transplanted to New York City’s East Village. We definitely didn’t speak the same language.
As luck or fate or just railroading would have it, our high-speed York to London express developed a pantograph problem shortly after departure, and the computer or the dispatcher switched the train from the high-speed “down” (or was it “up”?) track to the low-speed “down” (or was it “up”?) track so the maintenance personnel could have easier access to the equipment and not tie up, or tie down, the high-speed down, or up, track. I agreed with that. So far, the computer and I were in perfect sync.
I looked at the timetable I had picked up in the station and saw that there was a second high-speed train that was scheduled to depart York at 0825 hours, and I felt even better. With any luck at all, the pantograph problem would get corrected, then the 0825 train could pick up our schedule, and we could assume its schedule and still get to London with time to spare. That’s how I’d do it, so naturally. I knew that would be how everyone, man, woman, or machine would do it.
Sure enough, the problem with the pantograph was corrected, and, now 25 minutes late, we started to move south on the low-speed down, or up, track. I figured we would stay on the low-speed to the next interlocking, where we would be lined behind the 0825 onto the high-speed down, or up, and everything would be, as they say in England, “tickety boo,” or “Bob’s your uncle,” or some other expression just as wonderful. I was running through the list in my head of British slang, when the 0825 zipped by us at every bit of 200 km/hr.
Feeling ever so validated of my universal ability to determine the right course of action on any railroad anywhere and at anytime, I proudly told my British colleague what was going to happen and how we would run on the second train’s schedule. Bob’s your uncle indeed.
However, someone, somewhere, somehow apparently wasn’t concerned with my knowledge of British slang, or how I thought the railroad should run, because our train did not cross over to the high-speed down, or up, track at the next interlocking, or the interlocking after that, or the interlocking after that.
We were on the low-speed down, or up, and there we stayed.
We were 47 minutes late by the time we were routed back to the high-speed track. I was, as you might have guessed, fuming. The high-speed had remained clear, without another train to London passing us, just as the timetable said.
Finally, I looked at my British colleague/host, a soft-spoken man of unfailing courtesy, kindness and ability, and said: “Graham, whose decision is this to keep us on the down low, or low down or whatever, the computer’s or the dispatcher’s?”
“I don’t know,” Graham replied. “Why do you ask?”
“Because if the decision was the computer’s, when we get to London, I’m going to find it and unplug it.”
“And if it was the dispatcher’s decision?”
“When we get to London, I’m going to find him, or her, and I’m going to fire him.”
So here we are, on a platform that’s supposed to be almost intuitive—so easy a child could handle it. And here I am, without a clue, frustrated more than any child, having to figure out how to move and archive all the other postings on the other platform to this one.
Piece of cake, easy as pie, absolutely scrummy.
David Schanoes is Principal of Ten90 Solutions LLC, a consulting firm he established upon retiring from MTA Metro-North Railroad in 2008. David began his railroad career in 1972 with the Chicago & North Western, as a brakeman in Chicago. He came to New York in 1977, working for Conrail’s New Jersey Division. David joined Metro-North in 1985. He has spent his entire career in operations, working his way up from brakeman to conductor, block operator, dispatcher, supervisor of train operations, trainmaster, superintendent, and deputy chief of field operations. “Better railroading is 10% planning plus 90% execution,” he says. “It’s simple math. Yet, we also know, or should know, that technology is no substitute for supervision, and supervision that doesn’t utilize technology isn’t going to do the job. That’s not so simple.”