Graduation Speech 2017
Duke Univeristy
Department of Computer Science

© 2017 Ronald Parr

Welcome and congratulations to parents and graduates! I'm delighted to see that we are sending so many Duke computer scientists out into the world. This is a great time to be expert in computer science, not just the person who answers "Android or iOS" for the rest of the family at Thanksgiving, but certifiable, real-deal experts. It's printed on your degree, so it must be true!

Unfortunately, it turns out that I have some bad news for you: Expertise is dead. At least that's what the experts say. There's a very interesting and troubling essay and book by Tom Nichols called, "The Death of Expertise." Nichols's concern, and that of many others, stems from an odd paradox of our times that has largely been brought about by the efforts of computer scientists: The sum total of human knowledge is now LITERALLY at our fingertips (hold up iPhone), yet we don't seem to be making better decisions. Too many of us fancy ourselves to be supremely skilled at navigating the sea of information and misinformation that is before us, and we de-value the accumulated knowledge and wisdom of true experts.

I will offer two examples.

Example #1: In 2008, a model named Jenny McCarthy appeared on the Larry King Show to argue that vaccines can trigger autism. She was just one of many self described experts at the time whose advice put many at risk of a public health crisis.

The roots of the vaccine-autism myth can be traced to a discredited and retracted 1998 research paper. McCarthy, who possesses only a high school degree, was subsequently asked about her qualifications to be making scientific claims. She replied that she got her degree from, "The University of Google."

Of course, a true expert would understand the fundamental implausibility of such claims, and would be aware of the chain of events that led to the unraveling of the purported vaccine-autism link, a chain that began well before McCarthy's appearance on Larry King.

Example #2: As most of you know, the United States Department of Energy is a cabinet-level department that has many responsibilities which include both nuclear energy and nuclear weapons. In 2005, George W. Bush appointed Samuel Bodman, who earned a Ph.D. in chemistry from MIT, to head the department. He was succeeded by Steven Chu, a Nobel Laureate in physics. Chu was succeeded by Earnest Moniz, who earned a Ph.D. in theoretical physics from Stanford. Moniz was succeeded by our current chair, Rick Perry.

Perry could not remember the name of the Department of Energy in a 2011 presidential debate, but perhaps that's OK because it was one of three departments he said he planned to abolish anyway. Perry's highest earned degree is a bachelor of science in Animal Science from Texas A&M. Texas A&M is a respectable school and his degree could have included relevant coursework, but media reports of his transcript reveal D's and an F in chemistry classes, a D in Economics, and a D in Trigonometry. Many students find some of those classes hard, but Perry also earned a D in a class called, "Feeds and Feeding," and another D in a class called, "Meats." Perhaps none of this matters though because, like Jenny McCarthy, he can get a degree in anything he wants from the University of Google and we can be confident that by now he is as expert as the Ph.D. scientists who preceded him.

How did we get to the place where we take medical advice from models and energy advice from people who seem unqualified to compute the calories in a ham and cheese sandwich, or even remember all three ingredients?

Perhaps if these were isolated incidents, it wouldn't be so troubling, but these are just a couple of examples of a larger crosscutting trend in society - so don't feel smug just because of your political affiliation. We, nearly all of us, it seems, are increasingly comfortable ignoring established expertise and aligning ourselves with the messages, or the messengers, we find most appealing.

That's the bad news. And it is indeed very bad. But there's very good news too, and it takes two forms. First, I'll explain how your education might help you avoid the trap of ignoring expertise, and then I'll explain what you can do to help others.

Now, you might be thinking wait a second... The chair of the CS department just told us that we're all experts, so what's the problem? Each of you has achieved some level of expertise in your major, and possibly a little expertise in some other areas. That's no small accomplishment, but the most important thing you can get out of a good education is not pride in what you know, but humility in the face of what you don't know.

In your four years here, you have built your expertise using a wide range of experiences and resources: You wrote code. You debated with friends. You read books and academic papers. You went to office hours. You searched the internet and, yes, some of you used stack exchange. All of these experiences built a foundation of knowledge, but also a tool set for expanding your knowledge and a set of strategies for testing your knowledge. The resources you will have for expanding your knowledge after graduation are immense. More than any previous generation, you have the potential to live every professor's hope for his or her students - a lifetime of continued education and intellectual stimulation. But you will build towers of expertise and wisdom on your Duke foundation only if you are humble. You must be willing to understand where your foundation is weak, or where you need to extend your foundation to support more. You will need the help of experts in other fields and you should respect the investment they have made in THEIR foundations. Of course, this doesn't mean that you should accept expertise uncritically - you should always question and seek justifications because that is how expertise is tested and how new expertise is built.

So this is the first bit of good news: People with your education have unprecedented opportunities for a lifetime of learning and accomplishment - if you play it right.

If, however, your education becomes a source of hubris rather than a source of humility in the face of new challenges, then you will not build towers of knowledge and wisdom; you will build houses of cards. Sadly, this is where we find ourselves today as a society. We take medical advice from celebrities rather than doctors. We make policy decisions based upon what feels good, rather than facts. Our public discourse often devolves into anecdotes and puffy ad hominem attacks rather than a search for the truth. Everybody fancies him or herself an expert, but few respect or are guided by expertise.

And this brings me to the second bit of good news: Nobody is better prepared to change this than the people in this room. The computer scientists who came before you gave us the devices, the protocols and the algorithms that make the near instantaneous access to data that we enjoy today possible. Now it is your turn to change the world by giving us instantaneous access not merely to data but to knowledge and to wisdom.

We already see the beginnings of how this can happen. Google and Facebook have started projects to combat fake news. Just in the past few months multiple projects have started to automate fact checking. These projects draw upon the skills of computer scientists in databases, machine learning, and user interfaces to make it easier for people to tease apart curated knowledge from rumor and misinformation. The skills of computational social scientists will, no doubt, come into play as we further search for ways to incentivize truth-seeking behavior. Your opportunities to use your talents as computer scientists to do good are enormous.

But our skills as computer scientists won't get us all the way there. Our culture's increasing disrespect for expertise stems from impatience - it's a retreat into what is comfortable and easy, rather than a willingness to invest in learning and in people. No amount of automated fact checking alone can fix that, so here you must draw upon your softer skills as Duke graduates and lead by example. When the Thanksgiving conversation switches from "Android vs. iOS" to something more awkward and more consequential, you'll need to remember to take your expert hat off, and find the grace and humility to be a student once again. Don't overestimate what you know, and don't underestimate what you can learn if you're willing to listen. Sometimes the act of listening accomplishes more than pulling out the phone at the dinner table and delving into the ritualized contest in confirmation bias known as Google-fu.

So, graduates, we send you off to a world shaped by computer science, a world shaped by people very much like you. Now it is your task to use your talents both as computer scientists and as human beings to make it better. I have no doubt that you will, and that you will make your professors and your families proud, as you have already done these past four years.

Congratulations and best wishes class of 2017!