Friday, January 4, 2013

Highly Specialized Roles

[Approximate Reading Time: 5 minutes]
[Mood: Questioning]

Take a look at these word formations -

"Visual Depth Perception from Texture Accretion and Deletion: A neural Model of Figure-ground Segregation and Occlusion"

"Characterization of the Nitric Oxide and Adrenomedullin Signaling Pathways in Normal and Diabetic Mouse Retina"

"Development and Applications of Diels-Alder Cycloadditions of 2’-Hydroxychalcones"

The above are dissertation titles of the 2012 PhD graduates of Boston University. There are more tongue twisters on this site for your reading pleasure.

I wouldn't be surprised if I meet a stranger on the street and find out I have no clue as to what they do all day at work. Who talks about work these days anyway in a social setting? It's understood that work is boring, at least definitely not as much fun as the party you're in. I could ask someone "how's work?" and expect "ah, you know, the same old". That pretty much ends the conversation about work and I know that if I press anymore, it's like dragging a horse where it doesn't want to go. But I digress.

It appears certain types of work, those generally referred to as professional work, is overly narrow and specialized. I have a hard time explaining what I do all day to my friends in any significant detail. And some of these friends are in the same general field as I'm. You know what I'm talking about if you ever felt like you needed to come up with a good description of your day-to-day job in two or three sentences that would make sense to most people.

It takes up to 11 years of post-graduate education for one to become a "specialist" in the medical field. To be sure, we all want only the most qualified and highly trained doctors attending to us and our loved ones but do we really see any correlation between the length of study and a doctor's abilities? Medical school was never as specialized as it is today.

Specialization has crept into almost every profession. Computer Science and Electronic/Electrical Engineering used to be one and the same department in universities. Today, Computer Science is a large enough field in its own right that leading technology companies often look for either research or work experience in further sub-fields of Computer Science, like Artificial Intelligence, or Machine Learning.

Why is this important? Specialization keeps us from seeing the big picture. Used to such ever narrowing fields of study, we tend to lose perspective of the whole. What does the study and practice of Machine Learning accomplish in the bigger scheme of things? What kinds of products does it enable? What kinds of effects do those products have on society and culture? If you cut through the propaganda put out by technology companies on how their products are changing the world, you might notice the utter lack of accountability at the management layers above you. The fact is, technologies are deployed by corporations not to make the world better but to expand their bottom line. But having been highly trained in one field or another and working professionally in one role or another, we tend not to see this. We find it hard to see beyond our immediate day-to-day work schedules and project timelines. We fail to see how they're affecting the world at large. We fail to connect the dots.

Aerospace engineers spend years learning about flight. Throw in a few more years of post-graduate research and you might very well be an expert on the landing gear of mid-size aircraft or recent developments in door design of large aircraft. When someone is as specialized as that, designing an aircraft door becomes a full-time job. The issue one often misses is this: it's not always easy to tell if the door will end up on a commercial jetliner or a military cargo plane. I think a responsible aerospace engineer would want to know how his or her work is being used or going to be used in the future. To be sure, there are plenty of aerospace engineers who would jump at a chance to help design the next generation killing machine but in most fields, from medical research to software engineering, thousands of scientists and engineers are toiling away day and night working on new technologies or algorithms that are very likely to be put to uses they wouldn't approve of, if only they knew. At the end of the day, your corporation or research lab owns your output.

The system is set up so that we all work on highly specialized tasks that, when put together, have enormous potential for impact, both positive and negative. History tells me most of us inadvertently get sucked into helping the war machine, or suppression of human rights, or wanton destruction of the environment by the powerful. Mind you, the powerful know exactly what the end products of your toil would be. If they don't know yet, they will as soon as they figure out how to make money off of them, or to control society. They're running the show, and they're not telling us what the props are.

The next time we get excited about Big Data or massively parallel computing, let's try to expand our imagination and think of all the potential harm those technologies might do to people like you and me in the years ahead. Face recognition sounds rather cool until you realize that it's your teenager's face that the authorities captured on the security camera the other day. Big Data is an interesting problem indeed and we all love challenging problems with potentially innovative solutions. But can you make sure not to let a government official or a corporate sales person see blow-by-blow data on your wife's shopping habits or your husband's addiction to the NFL, data that he can slice and dice any way he wants.


Now, I know there are many who are aware of the implications but aren't as concerned as me. There are also those who think it's actually a good thing to be solving problems and building new technologies despite any risks, however large. There's no need to be so paranoid, they say. Yes, there's some risk of misuse, but we will be creating a better world overall.

I'm not so sure!

No comments:

Post a Comment