Augmentation, Automation, and The Future of Work: The Tom Davenport Interview

Tom Davenport’s Twitter bio describes him an “analytics person.” As the President’s Distinguished Professor of Information Technology and Management at Babson College, the co-founder of the International Institute for Analytics, a Fellow of the MIT Center for Digital Business, and a Senior Advisor to Deloitte Analytics, he might be slightly understating his credentials.

In a recent column in the Wall Street Journal’s CIO report, Tom argued that “Automation Is So Yesterday,” and that “augmentation — combining smart humans with smart machines — is a better strategy than automation” because business processes evolve too quickly to completely automate.

We found Tom’s argument especially compelling in light of the way our Wordsmith platform works for generating stories from raw data. The stories themselves are indeed automated, but only after smart human writers configure the platform. All told, the platform augments what a human employee can do.

Your recent column says that “Automation Is So Yesterday.” If I go back in time five years and read you that column, are you surprised? Have your views on automation and augmentation changed, or is this a theme you’ve been pursuing for a while?

I started some new research in this area about a year and a half ago. At the time I was more focused on the impacts of automation for knowledge work. But as I talked to companies and vendors about this, I became convinced that augmentation is both a preferable and more likely outcome in the vast majority of situations.

Do you find many people have an either/or, all or nothing view of automation? If so, why is that, especially when the evidence of everyday life suggests that progress typically happens through the blend of augmentation?

Yes, I think that automation is much more dramatic than augmentation. So you have a lot of smart people–Bill Gates, Elon Musk, Stephen Hawking, etc.–warning the society about automation and its impact. Augmentation will also have some impact on jobs and work, but it will be much more on the margin. If these smart machines are not going to become our robot overlords, it’s a little less newsworthy.

What are a few promising examples of augmentation that most people might not be aware of yet?

In personal financial advising, you have this idea of the “robo-advisor” that sounds very automation-oriented. But in several firms it’s actually an augmentation approach that combines human and online advice. Certainly all of the IBM Watson oncology advisor systems are very much augmentation-focused; nobody thinks that oncologists will go away. And of course the news-oriented systems that Automated Insights offers have not put any journalists out of work as far as I know.

It seems like many employers and employees operate under the assumption that the jobs of the future will look pretty much like the jobs of the present. What would happen if employers and employees truly embraced the idea of an increasingly machine-augmented future?

I have co-authored a book about this with Julia Kirby. It’s called Only Humans Need Apply: Winners and Losers in the Age of Smart Machines and will be out in a couple of months. It is about several key ways in which smart humans and smart machines will work closely in the future. Of course the specific work design will vary by the specific job and task.

One current augmentation/automation debate is around self-driving cars. What advice would you give to car companies seeking fully autonomous vehicles?

I do hope that at some point–maybe when I get too old to drive–that there will be fully autonomous vehicles that can be summoned when needed. But I’m pretty sure in the short run that we’ll be looking at augmentation, and in fact that is required by some state laws. I understand why Google, for example, is focused on pure autonomous vehicles–it’s easy for humans to stop paying attention when there are some autonomous capabilities–but I think they’d make faster implementation progress with a human in the mix.

A skeptic might say to you, “sure, smart humans are still better at many tasks for now, and so augmentation is still better for now. But there will come a point where machines get smart enough that they’re better than human an any activity.” How do you respond?

I think we would be foolish to discount this possibility. But it’s not going to happen anytime soon. The singularity is not near!

A version of this interview originally appeared on AutomatedInsights.com.