How Open Source Tools are Changing HR Research
I’m finishing the qualitative research for my annual report on the state of intelligent tools in HR tech. The research covers products and services claiming to use artificial intelligence in people analytics, recruiting, learning, talent management, payroll and HR information systems. I’ll discuss the results at the HR Technology Conference this fall, but here are some initial impressions.
Some things have changed dramatically. In particular, the methods of conducting HR research have changed from an unbearably slow peer-reviewed academic process to a venture-financed small laboratory model. The old approach allowed information to be shared by teams focused purely on advancing the science. The new way focuses on accumulating intellectual property while trying to achieve revenue goals.
It’s a big difference. Yesterday, the ponderous pace of academia produced foundations that everyone could build on. Today, small, private, revenue-focused labs push products out the door faster, but they don’t share information.
The large-scale providers of cloud processing and storage (Amazon, Google, Microsoft and Oracle) all provide some level of open-source tooling. Most of the so-called AI projects in and out of HR tech build on this backbone. The problem comes with results and processes specifically in the HR tech domain.
Also, when one organization using these tools learns something important, its leaders have little incentive to share it. For instance, more than a few companies have stopped using flight risk forecasting because in practice it leads to higher attrition. With no way to share that information and no shared notion of the future of the HR tech industry, information languishes while tech providers profit from reinventing the wheel.
This creates a perfect setting for the open source principles of transparency, collaboration, meritocracy, community and frequent release. A central facility for housing discoveries on HR Tech might accelerate industry growth. One company exploring this approach, Shaker International, home to Virtual Job Tryout, is the HR tech industry’s champion of open source.
One way of thinking about open source is as the intersection of commercial and academic styles. It’s intended to move the field forward through intense sharing of learning and techniques. Shaker International doesn’t believe that its advantage comes from hoarding what it learns. Rather, it shares its discoveries widely, staying ahead by helping the companies chasing it. That, in the eyes of Isaac Thompson, one of Shaker’s data scientists, is how you accelerate the growth of science.
Shaker, HireVue, Pymetrics and DeeperSense are just a few of the firms turbo-charging IO science with intelligent tools technique. Employers including Walmart, USAA, Facebook and Google field substantial industrial organizational psychology teams searching for competitive advantage for their workforces.
Thompson along with Nick Koenig of Shaker and Mengqiao (MQ) Liu of Amazon wanted to bring the rest of the IO industry up-to-date in the fastest, most thorough way possible. They wanted to get IO psychology focused on the use of machine learning and natural language processing. In 2018, they took over the annual Society for Industrial and Organizational Psychology hack-a-thon and refocused it on creating usable code and models for the industry in 2019 and beyond.