HR has a significant role to play in the emerging digital work environment, and HR professionals must empower themselves with sufficient knowledge and understanding of developments such as machine learning to effectively guide and manage this process, writes Rob Scott.
I’m a strong advocate of the aphorism ‘knowledge is power’ from a positive perspective of building intuition and ability to contribute to new thinking, innovation and creativity, rather than the negative connotation of control over others. And while continuous education and learning is an absolute necessity in today’s agile work environment, learning something completely foreign to your educational framework or work experience (such as machine learning) is daunting, to say the least.
For many HR professionals, the emerging digital work environment is shining a giant spotlight on their digital and technical skill/awareness void. What makes it difficult to rectify is the fundamental differences between a social science-based education, which most HR professionals emerge from, and a STEM-based education which underpins information technology and data science jobs.
Over the next few years, the influx of advanced technologies such as RPA, robotics, bots and machine learning (AI) capability will continue to change how we work, how we respond to business challenges, how we analyse and make decisions. Together with the realisation that technology is not going to displace humans in the short term, but rather augment what we do, it’s become obvious that HR professionals must supplement their skill set to effectively operate in a digital and AI world.
“For many HR professionals, the emerging digital work environment is shining a giant spotlight on their digital and technical skill/awareness void”
Some may resolve this problem by hiring data scientists and architects into the HR function rather than upskilling current staff. There is nothing wrong with this approach, however having ‘STEM’ educated resources focussed on e.g. HR analytics, reporting to HR leaders who have little in common from an education or appreciation perspective is likely to create short and long-term problems.
Just as today’s HR professionals learn ‘finance for non-financial managers’ which promotes common understanding, insight and the ability to engage in meaningful discussions and decision making of the financial kind, it is equally important for HR professionals to put aside any concerns and misconceptions about learning a STEM-based subject. Having the right insight and understanding of data models, how machines learn, types of issues, risks and opportunity, strengthens ones’ position as an HR leader and empowers you to get the most out of your STEM educated staff.
As a person with a social science background, I decided to put this to the test and enrolled myself onto a free Google ‘Machine Learning (ML) Crash Course’. It’s a 14-hour online self-learning course which includes some technical programming. Here are my key observations and learnings:
- In hindsight, it wasn’t as difficult as I thought, although I did feel completely out of my depth initially – but I pressed on. I did opt out of trying to fully understand and remember all the maths or completing the programming tasks. However, I spent quality time understanding what the formulas and programs were aiming to achieve. As I progressed through the course, I found myself recognising maths terminology and began understanding why the equations were important.
“Current HR professionals should urgently seek out basic training opportunities to build their insights”
- Most of my resistance to learning maths was preconceived and hinged on less-than-favourable school memories. It is possible for old dogs to learn new tricks.
- I found understanding the ML concepts easy, and the way the course is designed (video, support notes, practice, test etc.) supports adult education practices. I feel confident having a conceptual conversation about machine learning, framing an analytic outcome, the importance of data types and sources, validation, training and testing.
- However, what I really learnt is that there is no such thing as AI … It’s all clever maths, but there are also many reasons why a machine learning algorithm could be incorrect or biased based on a variety of mathematical assumptions as well as individual personal perceptions. Knowing the basic risk factors has empowered me to ask the right sort of questions.
- HR professionals and data scientists need each other for digital outcomes to be successful. It became obvious to me that the skill profile of a data scientist does not lend itself to ask the right HR type questions. Just as HR professionals need to learn the basics of machine learning, data scientists need to learn the fundamentals of HR to promote meaningful discussion, decisions and beneficial outcomes.
- There is an urgent need for AI/machine learning courseware specifically designed for non-data science leaders. While this course was good, it’s not ideal for the general HR population. We don’t need HR professionals to become data scientists, but they should have knowledge which enables confidence and full involvement in conceptual digital and machine learning design.
HR has a significant role to play in the emerging digital work environment. The more we immerse ourselves in understanding concepts of new technologies, the greater our value offering will be.
3 key takeaways: machine learning for HR
- HR professionals must empower themselves with sufficient knowledge and understanding of technology in general, data principles and machine learning concepts to effectively guide and manage future HR teams.
- Hiring STEM-skilled resources into an HR function where the HR leadership team is not conversant with machine learning concepts and principles is a significant risk which could lead to retention issues as well as flawed outcomes.
- Formal social science education courses are adapting their programmes to include appropriate STEM elements, but current HR professionals should urgently seek out basic training opportunities to build their insights.