To some, it might feel like the dog days of summer. For those of us involved in the 20th Annual HR Tech Conference, it’s planning time. Kicking off our pre-conference #HRTechConf series is an interview with industry analyst, John Sumser.
As Bill Kutik so aptly describes him, "John Sumser is the most original out-of-the-box thinker in HR and HRTech today.” For nearly a quarter of a century, Sumser has prodded, explored and chronicled the evolution of HRTechnology. He's an independent analyst who finds the edge of the HR conversation and explores it. You can find his work (and a treasure trove of industry reports) at hrexaminer.com. Recently, he launched a weekly newsletter of useful links and meaningful commentary called the HRIntelligencer. Its goal is to help vendors and practitioners stay on top of the fast changing world of machine led decision making.
Currently, John is in the middle of deep research into AI for HR. Preparing a Market Landscape report based on interviews, demos and research on AI trends in general and specific looks at 25 companies in the HR sector, he’s looking at all of the HR silos. The report, which will include practical things for HR professionals to do, is slated for availability on October 1, 2017.
John, as one of the HR Tech industry’s most respected analysts, you’ve taken an early interest in the critical role AI will play in the workforce. Can you tell us what HR professionals should be aware of?
It’s the earliest of early days. Maybe you remember the earliest “horseless carriages?” Well, you probably don’t remember but I’m sure you have seen pictures. Those early bits of technology involved trying to take the old problem and solve it with the new stuff. That’s what most of AI in HR Tech looks like today.
But, there are really, really smart people doing the initial experiments. It used to be the case that R&D was completely distinct from the cash flow parts of technology companies. Today, R&D has to pay bills as well as explore. So, a good number of the emerging vendors are solving current problems while staring deeply into the future.
Using AI will pose interesting new challenges for HR professionals. Your computer is going to start telling you what to do. There is a host of ethical and leadership issues involved. For instance, what do you do if you disagree with the machine’s insight, guidance, direction or decision. There are enough big questions that every HR professional should educate themselves on Machine Learning, Artificial Intelligence and Predictive Analytics.
Over the last 10 years, the drumbeat has been pretty steady about HR data and analytics. Is AI just another iteration or are there distinctly new business outcomes that we can expect from it?
Some things have really changed.
a. The open source movement makes it possible for all tech companies to get a running head start on the machine led decision making process. There is an array of tools available to help with the building of niche specific applications. In other words, earlier generations of software development required the early adopters to make it all by themselves. That is no longer true. As a result, innovation reaches us faster.
b. The failures of the data and analytics juggernauts helped us see the problems. We now know that most companies have horrible data that needs to be cleaned, scrubbed, tagged and simplified. The early adopters all learned this when their analytics projects failed because you can’t create standardized reporting from a nest of tailored workflows. Analytics always requires standard processes to measure.
c. The early analytics catastrophes helped clarify what is required. Today, we face an explosion of data. Every day, it’s something new. Companies like SwoopTalent are pioneering the development of tools that identify data anomalies and package them for decision making.
d. As predictive tools evolved, we learned more and more about emerging software liability problems. Operations like Engage Talent source their data from outside the client in order to eliminate those concerns.
e. Machine Intelligence will be everywhere. In the early days, automation will be relentless. Any repeatable process will be incorporated into the machine. The analytics people believe that everything that matters can be measured and managed. It turns out that that’s a great way to describe all of the things that can be automated. We may look back at the analytics movement as the first, necessary step in something much bigger.
That said, real value is created in processes that are not standardized (yet.) Employment will increasingly involve work on things that don’t solve readily.
There’s been a lot of noise about bots. Are they truly AI or strictly transactional algorithms?
I wish that distinction was easy to make. Is Google Maps smart or is it just a transactional algorithm? In my opinion, fast transactional systems are difficult to distinguish from AI. I’m hearing stories about candidates trying to date the chatbot who helps them with their job application.
The funniest paradox about intelligence is that we generally believe that we are competent to distinguish forms and levels of intelligence. Just consider the American political conversation. One side is very clear that there is no intelligence on the other. I forget which side that is.
All of this emerging technology solves problems by clipping out repetitive and redundant parts of a workflow. You can do that by setting chatbots loose or by identifying the next step in training based on the new employee’s temperament, prior experience and current psychometrics.
In other words, intelligence is a subjective opinion and everyone has their own. One of the first things we’ll see as this new landscape emerges is that we stop calling it intelligence. I prefer “machine led decision making.”