Article: Can machine learning help overcome human failures?

Technology

Can machine learning help overcome human failures?

A lot of HR technological tools today are incorporating machine learning to help, support and empower humans at their workplaces. Here is a glimpse of a few ideas shared at the People Matters and Microsoft India roundtable on the same subject.
Can machine learning help overcome human failures?

There has been a lot of hullabaloo over artificial intelligence and how it would change the world in the course of the next few years. People Matters and Microsoft India organized a roundtable which discussed key ideas on artificial intelligence and how it is impacting workplaces today. This article attempts to capture some of those key ideas and present to you in a simplified format.

Artificial intelligence is actually at a very nascent stage

Despite all the discussions that have been undertaken in many of the public spheres, there is still a lot of confusion about what exactly can artificial intelligence do for you. First, for any system to become intelligent, it will capture data, analyze it, learn from the frequent interactions with the user, and then intelligently respond to the situation. The end of the goal is for the machine or the software to take decisions without human intervention. This leads to automation in processes. 

An example, which was discussed during the roundtable, revolved around automating the process of reading hundreds of resume and shortlisting candidates dependent on an algorithm which ensured that only the best fit for the job was selected.

Is Machine Learning only restricted to the domain of HR

Machine Learning is constantly evolving and very differently across industries. To understand why something like this is possible, one needs to understand that the evolution of any technology, and especially something like artificial intelligence, the evolution is completely dependent on the use cases. The use cases determine the data that is captured, the algorithm which is developed to organize the data, and finally, the purpose for which the analysis of the data is put to use.

In one of the instances, Sanjoe Jose, co-founder, and CEO of Talview, during the roundtable gave an example of how an algorithm which determines the survival of the patient, referred as ‘survival analysis’ was adapted to create a similar algorithm which determines the time it would require for a position to get filled in the organization.

AI and L&D

Similarly, artificial intelligence is being used in the domain of learning and development, and which is not just restricted to skilling talent, but also managing and guiding talent within the organization. Consider, that most of these technologies are cloud-based, what is they further enable the employee to chart his own career path in the organization. The software would let him know of all the training and resources available to him to develop the requisite skill, which will help him fill his desired role in the future.

During the roundtables, it was extremely well demonstrated in the case of a recently onboarded employee who is looking for resources which would enable him to carry out a task. Organizations today understand that employees do not just learn on the job by default, but they also look forward to ‘learning’ on the job. 

In larger setups, individual attention and mentorship not always available and which is why for many large organizations it makes sense that a 24/7 chat bot communicates with the new joinee and enable him to perform on his job. 

The process is actually quite simple: once the data at the organizational level is collected and organized, it can simply be grouped dependent on the algorithm deployed. Further, when the chatbot receives a query, it is translated into the language that the computer understands, the query is processed and the solution is made available, this solution is then presented to the person who asked the query, in the language that the person understands.

For example: A person who wishes to organize an office offsite but does not know how to go about it, would simply type in the chat and ask for help. The chatbot would then give it a list of videos or articles which his colleagues or peers would have accessed in the past, and also the ratings and comments they had for that particular resource. Similarly, it could also connect that person with somebody who has conducted successful offsites and they could connect with one another over chat or on call, enabling the transfer of knowledge within the organization.

We need more data scientists

Sandeep Jayaprasad Alur, the Director Partner Technology Engagements, Microsoft India gave the answer to this question, when he revealed to the audience during the roundtable, that tech giants like Microsoft have been working on developing cloud-based artificial intelligence based platforms since the last decade, and have only recently unveiled the technological platform, because it is only very recently that organizations have had access to humongous data that they have today.

The infrastructure which consists of the cloud computing, algorithms and the abundance of data is today readily available to organizations, but what is lacking today are the people who would take these technologies to the next level, and by people, it specifically meant the data scientists. He defined them as people who knew much more about statistics than a programmer, and much more programming than a statistician.

Using Machine Learning for engagement with candidates

A major challenge that the recruiters face today is responding to the hundreds of applications which organizations receive. The recruiter finds it difficult to effectively engage with all of them. The challenges start with analyzing the data in the resume, to engaging with candidates on an individual level, arranging for assessments tests and interviews, lacking data inputs which would enable the organization to make the best decision.

Today, all of the above is being taken care of technology which are an integration of data capturing, analysis of the same within the domain of the algorithm deployed, mapping the skills to the job roles, and uses a user-friendly interface (chatbots are most popular) to respond to candidates, and hence leading to effectively managing hiring.

From baby steps to shaping the future of tomorrow

During the roundtable, a question was posed to one of the speakers, that what most of what they show that technologies exhibit is actually extremely advanced. To which the speaker answered that they have the algorithms but what technology would lack on initial installation is the organizational data. But because the algorithms learn from interactions with the user, it would acquire the required data both at an individual and organizational in no time, and hence would then accomplish all the complex tasks (the example of a person wanting help with organizing an office off-site), very easily.

Hence, to summarize it is too soon for us is to completely imagine the altered employee experience of tomorrow. But what we do know for certain is that our employee experience at workplace will never be the same again.

Finally, can machine learning help overcome human failures

In most cases, the algorithms actually predict the probability of the desired outcome happening. Which means if the software tells the recruiter that of the hundreds of resumes received by the organization, only five are the best fit for the job, then that means that is the most likely outcome that it has arrived at. Most algorithms are predictive models which boast of a certain level of accuracy, but because machines do not err, in the sense that it won’t skip any of the resumes because it felt tired; machine learning to large extent would enable humans to perform to the best of their abilities.

Read full story

Topics: Technology

Did you find this story helpful?

Author

QUICK POLL

What is your top focus area for reinventing work in the hybrid world of work?

Can we take a more holistic view of how we reward people?

READ our latest issue for a broadened perspective.