�ݺ�ߣ

MACHINE LEARNING: AN HR PERSPECTIVE
by Prof. S. Chandrasekhar & Prof. Nina Jacob
There is a definite HR angle to machine learning. Consider the automobile industry
in America. Since the time of Ford, and until recently, assembly line workers were
the mainstay of this industry. Management experts expended considerable energy
through work studies, to devise efficient methods by which assembly line workers
could be at their productive best. According to “Robotics Online” (January 2017),
by the 1980s, billions of dollars was being spent by car manufacturers world-wide
on assembly line. As a result, assembly line workers were continuously forced out
of the work-force, as they were being replaced by smart machines. The job
requirement in the machine learning era was for technologists who could program,
run, and maintain the robots who were performing assembly line jobs.
JOBS IN THE MACHINE LEARNING AGE
The technologists who manage robots need to be trained computer programmers.
Their levels of education are higher than those of assembly line workers.
Consequently, their remuneration is higher. When assembly line workers choose to
return to university and equip themselves with degrees in computer programming,
they are upskilling. By upskilling, the former assembly line workers can oversee the
smart machines who have replaced them. Unfortunately, not all displaced assembly
line workers have the resolve and drive to upskill when past their prime. Such
workers can become disgruntled and embittered. These were the workers who
voted Donald Trump into office, in America.
Displaced low-wage workers are those who are negatively affected, as they do not
have the background or opportunity to upskill. This then creates a new type of
have-nots: the financially challenged, low-wage workforce displaced by automation.
This exacerbates an existing inequality which should be addressed by HR
initiatives at the government-level. According to the Indian 11t h
five-year plan, only
10 % of the Indian workforce has a university education. The remaining 90% have
no recourse if their jobs are automated and they are rendered jobless.
The new have-nots will then get chained to a life at the bottom of the pyramid.
They have spent all their work-lives in routine, repetitive jobs. Such jobs are the
ones which are easily replaced by smart machines. Martin Ford has pointed out in
his gloomy book “Rise of the Robots’ (2015) that the only way traditional workers
can survive in this age of machine learning is by switching from routine, unskilled
jobs to non-routine, skilled jobs. Meanwhile another category of jobs is now on the
threshold of becoming redundant on account of machine learning, that of white
collar workers.
The crying need of the hour is for the government to pass legislation that ensu res
that companies upskill those employees who are going to be replaced by smart
machines. The law passed in 2014, requiring companies with net revenues of
greater than ten crores to invest 2% of their revenue on corporate social

responsibility activities is beginning to do wonders. We now need similar legislation
for companies displacing jobs through automation. An article published in “The
Hindu” (July 4, 2016) notes that the Indian textile industry is likely to generate only
29 lakh jobs for humans in the next 5 years, as opposed to the government target
of 1 crore new jobs in this industry. This is because automation is sweeping away
jobs in the textile industry just as it is in the mining industry. This is bound to
create widespread unrest amongst the employees of these two sectors, as well as
volatility in Indian society at large.
AI/ MACHINE LEARNING AND HR
Artificial intelligence/ machine learning is taking off in a big way in almost all functional areas of
management be it finance, marketing, supply chain, logistics and HR.
At present though the adoption is somewhat slow in HR as compared to other functional areas
of management it is making inroads very fast.
Artificial intelligence/ machine learning are software programs that mimic the way humans learn
and solve complex problems. These systems are different from other application packages that
have been used in some of HR application in the sense that they learn and adapt to the new
environment as and when new data becomes available. In usual software packages the
outcomes are explicitly programmed and they do not have the ability to learn and adopt.
With adoption of technology like mobile, cloud and social media large amount of textual data is
generated specially in HR area. Previous computer technology was not equipped to handle
such large volume of textual data. With big data technology maturing it is possible to analyze
such huge volume of data and come out with behavioral patterns that were earlier not possible.
Some of the areas in HR where machine learning has been successfully applied are:
1) increased success rate in hiring by profile matching.
2) attrition detection
3) post hire outcome
1) Increased success rate in hiring
Without machine learning when selecting the prospective applicants from a pool of CV’s is
basically done by keyword matching. With machine learning and text analytics one can analyze
the data from other sources like blogs, posts in social media, tweets and retweets, posting in
professional social media sites. Such selection will be a better matching than the simple
keyword search.
2) Attrition detection
This is a very important HR function which most HR managers are worried about. Organizations
spend quite an amount of resources on training and development. If attrition is not controlled it
will be a big issue. Employee surveys are carried out to detect this trend. The survey has so
many dimensions it is humanly difficult though not impossible to detect hidden patterns which
are predictors to attrition. Machine learning can identify certain signals and combination of

signals which are difficult to spot by humans. This can aid HR in taking suitable corrective
action.
3) Post hire outcome
Another important HR function which is difficult to predict. By better profile matching at
recruitment stage using machine learning and tracking the candidate behavior in different
channel one can predict the outcome. Also using web analytics, one can track who saw the ad,
which websites are successful than others. Keeping track of interview process, one will be able
to find out which interviewers are able to spot talent better than others. Such data can be fed to
machine learning and train the machine to predict post hire outcome.
AI/Machine learning technology is maturing and we are sure in years ahead it will have
significant impact in HR functioning.
-------------------------------------------------------------------------------------------------------------------------------

�ݺ�ߣ

Machine learning 1

More Related Content

Machine learning 1