Why Technology Still Needs the Human Touch

Last week I had the privilege of attending the ITA 2014 Fall Collaborative in Washington D.C. This meeting was co-located with the Digital CPA Conference and we were privy to some great speakers. One of my favorites was Nicholas Carr, who has written extensively about the intersection of technology and human progress. A favorite breakout session was Marc Teerlink, the Chief Business Strategist for the IBM Watson project. Upon reflection, these two presentations were connected in more ways than I thought.

IBM Watson has moved from R&D to the commercialization phase. IBM is making a major investment ($1B) in this effort that uses cognitive computing to translate data into dollars.

Teerlink provided many examples where this is being applied by early adopters in healthcare, financial services and other areas that fit his framework of “observe, interpret, evaluate and decide.” Essentially, this refers to knowledge work and how to augment our ability to absorb the vast source of data that are available to us. He noted “we don’t have a data problem, we have a filter problem.” By this he meant that we so often feel overwhelmed by the volume and velocity of data that comes at us, but the real issue is that we don’t have a filter mechanism that tells us what is relevant for the situation.

Later, Carr shared his view of the new world of technology we live in. Essentially he stated that the cloud is our data center—a large central utility—much like the power plant of 100 years ago. Local computing (private data centers, local servers) is being displaced just like steam engines and local power plants were in their era. We are rapidly transitioning from the old infrastructure to the new. Our biggest challenge may be rethinking our business and anticipating what it is going to change. A key part of that is understanding where we add value and separating routine activities from innovative, unique and knowledge enhancing skills. One hurdle is integrating deep automation.

Now we’ve come full circle. Deep automation is based upon capabilities such as IBM Watson and highly sophisticated combination of technologies (cloud, mobile, big data, internet of things, social) that come together in ways we are just beginning to realize. For examples, look no further than driverless cars and 3D printing.

In Carr’s recent book The Glass Cage, he warns that there are unintended consequences of automation. At the most basic level, there are two categories:

• First, complacency. We become complacent because we trust technology to work flawlessly. We substitute the computer for our thinking.
• Secondly, accuracy. We believe anything presented to us through the pane of glass. Even though if it was on paper or spoken to us we might question it. The fact that it is from a computer and presented dispassionately, we believe it.

Here is a vivid example of the dual pitfalls of complacency and accuracy. A Seattle bus driver flawlessly executed the route presented to him by his GPS. He was complacent and didn’t give the route a thought. Even when presented with road signs indicating low bridge ahead, it didn’t register caution. The GPS (computer) presented the data and he proceeded despite additional signs, the GPS continued the route and his bus with 12’ clearance crashed into the 9’ clearance bridge. The accuracy of the GPS was absent an input (vehicle height) and constraint (bridge height). He trusted its accuracy implicitly and followed the bias that since it’s automated, it must be accurate.

Photo: Dan DeLong/Seattle Post-Intelligencer

Our challenge, then, is to focus our attention, stay alert and use these powerful tools to augment our abilities. As Marc Teerlink stated, “Computers don’t ‘predict.’ they present.” They present information and knowledge based upon rich sources of data. But they don’t have the intuition (tacit knowledge) that could be codified into a set of explicit rules. When we confuse this simple rule, we fall victim to complacency and an accuracy bias that is dangerous.

Our relationship with technology is far from perfect, but useful nonetheless. I much prefer the world where we can use sophisticated technology and allow the override by the highly trained professional when required.

Additional resources:

• By Nicholas Carr, All Can Be Lost: The Risk of Putting Our Knowledge in the Hands of Machines

• Masters of our own destiny, Marc Teerlink’s Ted Talk on IBM Watson