Are we too stupid to let Watson make us smarter?
- 14 February, 2014 16:00
There's a guy in every family who sends email to the world containing some note from someone else on the Internet with an amazing fact. In my case, the guy apparently believes anything negative about Democrats, even though the family regularly suggests he goes to Snopes before sharing this information and embarrassing himself. His latest note suggests that Bill Clinton had to be pardoned because he criminally avoided the draft, making him the only felon President. Like most of the other "facts", it was utter BS, but off it went. I'm left praying no one I know find out we're related.
Unfortunately, this isn't an uncommon problem. It even has a name: Confirmation bias. Some of the worst decisions I've made in my life - buying motorcycles, cars and even homes - happened without thinking. I put all my information into a bucket that confirmed I was right, even though I wasn't. It took me a few years to learn my lesson, but now I generally check and double-check a decision and often find that the path that looked the best wasn't at all because I hadn't stepped back and considered all the facts.
This is where IBM's Watson could make a great contribution. So many truly horrid business decisions, from acquisitions to product strategies, have resulted from a combination of confirmation bias and argumentative theory. (The latter argues that the biggest jerk at the table must win an argument, lest he lose status, and tends to favor not appearing to be wrong over actually being right.)
It's Human Nature to Shoot First, Ask Questions Later
We've accomplished a number of amazing things as a race, but to say we are flawed as a race would be an understatement. Most troubling is the tendency to reach a conclusion quickly, then look for reasons to defend that decision. This does a great deal of damage.
Years ago, on a Siemens competitive analysis team, we briefed executive after executive on why the current strategy was going to fail spectacularly. Through hard research, we convinced every executive that we were right. They'd return to Germany, but instead of getting to fix the strategy, the executives got reassigned, and a new believer was sent to us. This happened three times before Siemens fired us - and then burned $US5 billion, proving us right.
Siemens' strategy was based primarily on the theory that ISDN would replace Ethernet as the prevalent networking standard. The rest read like something so unbelievable it would probably get rejected as a script for The Office, yet this PhDs - and they mentioned their degrees often - not only couldn't see it but wouldn't see it.
My most famous (or infamous) situation involved Windows Vista. I constantly said it wasn't ready and would fail if something wasn't done. (A later lawsuit made all those emails public.) Microsoft thought I was crazy - until the product launched and crashed in market. I was hardly the only one who saw that train wreck coming. But everyone else not only didn't see the signs; they refused to see the signs. The result was a boom.
Watson Can Make Us Smarter (If We Let It)
Watson, as you know, is IBM's thinking computer. This week, IBM presented something called Watson Jr. that will make Siri look stupid by comparison. But it's the next generation of this product that I find fascinating, because Watson is being taught how to reason.
Watson will have access to all the information you have - and it won't have confirmation bias or a need to gain status by appearing to be right. In effect, this cognitive computing will give us the equivalent of a "Stupid Alert" and be able to tell us when we're doing something we will later very much regret. This may be its most useful feature - and it could very well save the world:
- I recall a Senate hearing at the close of the Iraq conflict during which one general said rebuilding the country would cost more than the war itself. He was laughed into retirement - but he was also dead right.
- Scientists build larger and larger particle colliders and put them through much more rigorous experiments than they were originally designed to handle. I don't know about you, but when it comes to the death of our planet, 1,000-to-one odds, even if flawed, don't seem very reasonable. It's not like we can relocate.
- Global warming is a contentious issue, and one side of the debate seems to be more correct than the other, but it would sure help to know which site is, in fact, right before we are all dead. (Granted, the problem pretty much solves itself at that point.)
Watson should make us smarter. That's my hope. But I'm reminded of my Siemens example. In the face of a global catastrophe, if Watson told us that popular opinion was wrong, would we turn it off, thinking, "Gee, that can't be right"?
It sounds like a plot from The Outer Limits - and I sure hope for the sake of us all that it isn't also prophetic. Watson could assure our future. I wonder if we're too human to let it.
Rob Enderle is president and principal analyst of the Enderle Group. Previously, he was the Senior Research Fellow for Forrester Research and the Giga Information Group. Prior to that he worked for IBM and held positions in Internal Audit, Competitive Analysis, Marketing, Finance and Security. Currently, Enderle writes on emerging technology, security and Linux for a variety of publications and appears on national news TV shows that include CNBC, FOX, Bloomberg and NPR.
Read more about big data in CIO's Big Data Drilldown.