How to Work with Very Smart Machines" Q&A with Tom Davenport

SUBSCRIBE NEWSLETTER

The latest insights from your peers on the latest in Enterprise IT, straight to your inbox.

Q&A with Tom Davenport

Will a computer replace you at work sooner or later? This has increasingly become a concern of knowledge workers who in the past thought they would forever escape the fate of factory and low-level office workers who fell victims to automation. How should knowledge workers cope with the rise of smart machines?

In Only Humans Need Apply: Winners and Losers in the Age of Smart Machines, knowledge work and analytics expert Tom Davenport and his co-author Julia Kirby (a contributing editor for the Harvard Business Review) offer optimistic, upbeat and practical answers to this much-debated question. “The upside potential of the advancing technology is the promise of augmentation—in which humans and computers combine their strengths to achieve more favorable outcomes than either could do alone,” they write.

The author of Competing on Analytics (and 16 other books), Davenport is the President’s Distinguished Professor of Information Technology and Management at Babson College, the cofounder of the International Institute for Analytics, a Fellow of the MIT Center for Digital Business, and a senior advisor to Deloitte Analytics. Davenport, a Straight Talk Featured Blogger, recently talked about Only Humans Need Apply with Straight Talk editors.

You are widely known around the world for your work on analytics. Is this book a new area of research and writing for you?

It’s a new phase in a natural evolution of the analytics work that I’ve been doing over the past twenty years or so. This is true also for organizations—those that have invested in developing their analytics capabilities are well prepared for this new era where the analytics are embedded and invisible and frequently automated in many cases through technologies like machine learning. Whether you call the new approaches “deep learning” or “cognitive technologies” or “artificial intelligence,” they are all based on sophisticated statistical analysis. Recent advances in machine learning in areas such as image identification, text analysis and natural language processing are based on the progress that has been made in big data analytics.

Is knowledge work going to be completely automated?

Knowledge work was always going to be the great refuge of people who were driven by automation out of other types of jobs. But with the recent advances in machine learning and artificial intelligence, it’s quite reasonable to ask whether knowledge work is the next thing to go.  My answer to that question is a very clear yes and no.  I’m like an economist in that regard…  I think it’s very important to remember that computer systems don’t take over entire jobs, they take over specific tasks, and it’s certainly true that many knowledge work tasks are at risk of being automated. 

So while entire knowledge work jobs may not be automated anytime soon, I think it’s fair to say that there’s no room for complacency at all.  Anybody who is a knowledge worker today or aspires to be a knowledge worker has to wonder, “What’s this going to do to my job and how might I make myself more employable in a future of very smart machines?” There are knowledge work jobs for which there are already a fair number of tasks being automated.  I earn my living in part as a teacher and professor, and there are readily available machines now that do a better job than I do at presenting online content and then determining whether a student is really getting it or not.  These adaptive learning tools can diagnose how rapidly a student is learning a particular subject and present less complex or more complex content very easily. Similarly, we see today smart machines successfully applied to tasks traditionally performed exclusively by lawyers, wealth advisors, architects, accountants, digital marketers and other knowledge workers.

The good news, and I sometimes wonder if people want to hear the good news because the bad news gets so much publicity, is that there’s going to be a lot of jobs for people working alongside smart machines. In addition, these smart machines could lead to pretty dramatic productivity gains which could finance things like re-education programs and redeployment programs.

The focus on tasks rather than entire jobs lead you to stress augmentation rather than automation.

I hope that this becomes an era of augmentation rather than automation. I think we will be far more successful and satisfied with the outcome of our automation efforts if we don’t really fully automate and if we look at it as an opportunity to work alongside these machines rather than have them replace us.

I think augmentation is more of an optimistic way of thinking about the rise of smart machines, but also somewhat more realistic.  For the foreseeable future, in many cases we’ll see smart humans and smart machines working alongside each other.  My conclusion from the Tesla fatal collision is that Tesla recommended an augmentation-oriented solution where humans are paying attention to what’s going on. But some drivers decided that, “No, this is an automation-oriented solution and I can watch a movie or even nod off a bit while the car is being driven.”  That didn’t turn out well. 

We still need to figure out the exact form of augmentation in the autonomous vehicle realm and in other areas, but the general idea is that people would complement machines and machines would complement people.  Humans would either work alongside with and oversee the machines or focus on the tasks where humans are still better than machines. There are still quite a few of those. 

You describe in the book five strategies for working with smart machines.

Yes, these are what we call options for augmentation: Stepping Up means moving a level above the machines and making high level decisions about augmentation; Stepping Aside means letting the machines in your field take over and choosing to pursue a job that computers are not good at, such as selling or motivating; Stepping In means monitoring and improving the computer’s automated decisions; Stepping Narrowly means finding a specialty area in your profession that wouldn’t be economical to automate; and Stepping Forward means becoming involved in creating the very technology that supports intelligent decisions. 

Selecting the right augmentation strategy is not a trivial task and often requires investments in learning and in acquiring new skills. But each one these options is an invitation to participate in the future of work: either add value to what smart machines do or have the machines add value to your work.

This is useful and practical advice to knowledge workers. What advice do you offer to enterprises? 

Automation is often sort of a race to the bottom from a competitive sense in that other companies in the same industry do it and it ends up leading to lower margins and less innovation.  Smart organizations will say, “We are really focused on augmentation.  We’re not trying to just eliminate jobs. We want our employees to work alongside the machines.”

Now, there are a lot of smart machine technologies out there, so picking the right one for your application and specific goals of your organization is quite critical in designing the division of labor between humans and computers. 

Successful enterprises will be augmentation-oriented from the beginning and will develop a strategy for employees to take advantage of augmentation.  It’s going to take a while for humans to pick which of these five roles they’re interested in so employers may want to give them some time to think about that and learn the necessary skills. When employers invest in augmentation, they create a work environment in which knowledge workers are empowered to do more, not asked to do less—and as a result more value accrue to them as well as to the enterprise’s customers, partners, and other stakeholders.