The Skynet is Falling!: Which Data Science Outsider are You?

"The Intern" is a movie about a retired septuagenarian workhorse from another era who lands an internship at a hip internet startup. Could he possibly have something to teach the brilliant, young entrepreneurs? You bet he does! He's Robert De Niro! Only 2 years before, a suspiciously similarly titled movie called "The Internship" posed a similar question. Two 40-something-year-old school salesmen interned at Google. Could the technology giant learn anything from two experts from an era of door-to-door palm-pressing, big smiles, and taking clients out for drinks and cigars? Hmmm.....

Both films play on a fear of slipping through the technology cracks, and the desire to remain relevant. I've had the fortune of working closely with business leaders at large companies for a long time, and I've been a zealous proponent of finding the right applications of data science to fit the needs of organizations, maximizing the value and minimizing the risk of their applications. In my experience explaining data scientific applications and limitations to many working at organizations large and small, the fear of dwindling relevance has cultivated at least three types of data science outsiders. Which one (if any) best describes you?

1. The Eternal Skeptic
Have you heard the latest Stephen Hawking warning about the terrors of AI? Or the tweet from Elon Musk about how AI is more dangerous than nukes? The Eternal Skeptic lumps all applications of data science together, and decries all as the harbingers of the downfall of humanity. The fear of AI isn't unfounded. It stems from the idea that humans will create a computer that out-thinks humans, which will in turn invent something that out-thinks itself, and so on and so forth, making humanity look like an adorable relic of an obsolete past. It's referred to as the "singularity," or "Humanity's Last Invention."

The Eternal Skeptic also loves pointing out every time an algorithm gets it "wrong." For instance, if he already read through the fifth book in a series, and an online retailer's recommendation algorithm recommends that he would like the third book in the same series, this proves to the Eternal Skeptic that AI is all garbage.

If this is you, here are two simple remedies: remember that not all AI is the equivalent of putting Skynet online and inventing a super-intelligent computer that will render your professional expertise obsolete. To the contrary, many organizational applications of AI and data analytics are perfect complements to your experience, intuition and knowledge. Learn more about the different forms of prediction modeling and data science applications to get more comfortable with how you can use the analytic information to improve your intuition and expand the range of your expertise. Second, remember these three words, "compared to what?" When you're evaluating the effectiveness of a new analytic system (or any system), be sure to know to what you are comparing it. Maybe the system gets it "wrong" from time to time, but what is the error rate of the way in which the same questions were answered before?

2. The Way-Too-Early Evangelical Adopter
Some data science outsiders swing way too far in the opposite direction of the Eternal Skeptic. They are generally afraid of being left behind, and feel insecure about their lack of knowledge. They know some buzzwords, (like the Dabbler, discussed below), and throw them around very readily. These are the self-proclaimed "futurists" who believe that AI is a panacea and immediate substitute for human ingenuity. Who needs doctors anymore when AI can do it better? Who needs business leaders when robots can make pretty good strategic decisions without bias or self-interest? These guys are way too quick to welcome our new AI overlords. They want their organization to over-invest in technology and implement everything, often without sufficient regard for the risks of implementation.

If this is you, remember this: there is a great little parable described at the outset of the book, "Superintelligence, Paths, Dangers, Strategies" by Nick Bostrom, about a flock of sparrows that decide to domesticate an owl and elect to table the question of the immense problems associated with that proposition, opting to cross that bridge when the problems materialize. It's OK to consider the problems that will likely arise down the road. A little skepticism is healthy. Not all AI is a brilliant new thing that deserves adoption. In fact, some applications are rather pedestrian and blunt. If you can't tell the difference because vendors all showcase their services with flashy videos featuring inspiring graphics and interactive data visualizations, you aren't alone. I recommend befriending someone you trust who can peek under the hood on your behalf. Also, remember that it's OK to not completely understand everything. You should ask questions and expect data scientists to explain what they are doing without jargon.

3. The Dabbler
As a data scientist, I confess I am most terrified of the Dabbler. These guys took an intro programming or statistics course or two in college, or started a Coursera or EdX course on machine learning but didn't finish it. They know enough to be dangerous. There is nothing wrong with self-directed learning; in fact, I'm a HUGE proponent of this, especially in this field. But Dabblers lack the humility that should accompany their true novice status, and instead champion methodologies they read about online. Their vocabulary is full of jargon, and they will insist on using techniques or methodologies because those are the ones they've seen or heard mentioned at a conference.

If this is you, please consider the possibility that this organizational survival strategy could backfire terribly if you get something wrong, which invariably happens. You are likely better off genuinely stating your limitations and continuing to learn, than posing as an expert on something you are not. One of my favorite things about being a data scientist is the need to almost constantly learn something new. The field evolves quickly and probably very few people are genuinely 100% up-to-date on everything. Try to embrace the "perpetual student" ethos. I believe it will serve you well in any field anyway.