In the movie Her, Joaquin Phoenix has a virtual affair with his lovely artificial intelligence (AI) program, voiced seductively by Scarlett Johannsson.
(Spoiler Alert: This ends badly for the human.)
In the new movie Ex Machina, a hot shot programmer falls for Ava, an AI machine sheathed in a humanoid body. Ava definitely passed the Turing test.
(Spoiler Alert: This too ends badly.)
Given that humans don’t seem to be doing well in their relationships with smart machines, I am growing uneasy about my infatuation with an emerging class of very smart machines, including Watson, IBM’s AI system capable of answering questions posed in natural language and parsing a gargantuan amount of information quickly to give you an answer. I was first smitten with Watson and his fellow smart machines when “he” beat Ken Jennings at Jeopardy.
Is this going to end badly for knowledge workers? For knowledgement management (KM)?
Cognitive computing (aka machine learning) is definitely The Next Big Thing in analytics. It has the potential to troll the exploding world of data (healthcare, traffic, crime, sales patterns, etc.), apply algorithms and analytics, so that we mere mortals to make better, faster and smarter decisions.
To find out if there might be a darker side lurking in this promise, I interviewed the quintessential Big Thinker/Big Idea man himself, Dr. Tom Davenport.
Tom has written or co-authored 16 bestselling business books and is one of Harvard Business Review’s most frequently published authors. He has been named one of 10 “Masters of the New Economy” by CIO Magazine and the third leading business-strategy analyst (just behind Peter Drucker and Tom Friedman) by Optimize Magazine. He is also a friend and on APQC’s Board of Directors.
Tom is working on a new book [co-authored with Julia Kirby] exploring how knowledge workers can be augmented by smart machines lest they be automated out of a job by them.
I asked Tom “What are the characteristics of knowledge worker jobs that are going to be at the most risk?” The second question, of course, is what do we do about it?
TOM: I think almost any job that’s done by a lot of people and costs a fair amount—that makes it economical for somebody to develop an automated solution for it. If I’m a lawyer, a lot of work in commercial litigation involves looking at documents and trying to decide whether each document is material to the case or not. Is there something incriminating in those huge amounts of emails that come spinning around the company?
That [the electronic discovery of material documents] has a name, and it’s called “e-discovery.” It turns out that computers are pretty good at capturing images of those documents. By analyzing the text through predictive coding, computers can make a pretty good decision about whether a document is material or not without a human lawyer ever having to look at it.
Tom and Julia have come up with five steps that humans can take to augment smart machines, rather than be obsoleted by them: step in, step up, step aside, step narrowly and build a step.
If you don’t want this to end badly for your job, read more about these five steps in my interview with Tom.
Clearly, smart machines could eliminate many high-end knowledge worker jobs. That's the dark side.
On the bright side, cognitive computing will create exciting opportunities for everyone who uses knowledge in their work, including KM. At a minimum it will certainly help KM with content curation and amalgamation and delivery of search results, which tends to be the Achilles heel of KM technology. And, absolutely, it will provide decision support.
Let's make our vision even bigger than that: working side by side with the business units, experts and analytics function, KM and cognitive computing can accelerate the "rate of insight." Insight combined with action will prove the winning combination for the future. KM needs a seat at that table.
APQC is excited to explore the intersection of KM and cognitive computing. To read more about my interview on the future of KM and cognitive computing, go here: Preparing for the Future of Knowledgement Management.
Tom is still writing and still looking for stories to tell and for people and companies who have been affected by knowledge worker automation. Let me know if you would like to share your story with Tom Davenport.
You can go to the APQC Knowledge Base to read more of my Big Thinkers, Big Ideas interviews.
To subscribe to the Big Thinkers, Big Ideas podcast on Itunes or on APQCPodcasts on Podbean.
You can connect with me on Twitter @odell_carla