Artificial Intelligence in the World of Human Resources

Written by Consultinghouse Blogger | Jun 4, 2018 7:00:00 AM

 

 

 

Artificial Intelligence in the World of Human Resources

 

The Kepler satellite was propelled in the March of 2009 with the sole purpose of discovering other planetary systems. Almost a decade after its launch, the telescope fitted on the satellite had accumulated light waves from thousands of possible planetary systems. And through the analysis of this data, almost 2,500 new planets have been discovered till date.

Interestingly, in December of 2017, researchers at Google and NASA trained a neural network - which is basically a computer system modelled on our brain and nervous system – from all the indications that resulted in the positive detection of planets. After that, they made the system look for similar data in the vast database of light waves that Kepler had accumulated over the years. Surely enough, with the processing power of an everyday desktop PC, the neural network was able to identify a new planet in the Kepler 90 system. Obfuscated in a subtle mesh of data, discarded by astronomers, was an entire planetary system. Something manual sifting could have never pointed to, a relatively straightforward neural network found with mere few hours of training and some lines of code.

This is one of the most noteworthy proficiencies of Neural Networks and Machine Learning; searching for hidden patterns in sizeable volumes of data. Using it, pharma industry analyzes the immense amount of data from more than 900 vaccines and drugs currently at various stages of testing for the cure of cancer, Microsoft is employing it to decode the chunks of human DNA responsible for our immune system. To add to it, augmented reality, facial and gesture recognition, and self-driving cars...and artificial Intelligence has stopped being just a field of academic value, as it becomes increasingly clear that it will be the epicenter of the next huge technological revolution to shake our world.

Building on this preamble and moving to the crux of the article, many companies nowadays have a sizeable palette of data pertaining to the life of an employee in their setup – from the selection process (CVs, interviews), to training and performance metrics, there is an abundance of raw data and information. A neural network can be trained using this information about the best-performing employees and then look for the repetition of such patterns in potential employees currently in the selection or vetting process. This improved analysis of information would be very helpful in making human resource decisions – applicants that don’t meet the criteria of the neural network can be discarded objectively, optimized algorithms can be put in place to adjust newly hired candidates to a position they will perform best in etc. Furthermore promotions can be easily decided by a combination of satisfaction surveys, coworker reviews, performance information and likes. All this isn’t just theoretical as well. Despite being a comparatively young area in HR research, there are already applications being developed to introduce big data analysis pertaining to human resources into the mainstream.

Furthermore, analyzed data on employees also has a potential to be anonymously accumulated from various HR departments via a subscription-based platform. This can then be centrally analyzed and be helpful in determining patterns - like employees that have a higher propensity for an early exit.

One of the pioneers in the field of HR data analysis is Google, which has formed a People Analytics team for the sole purpose of providing problem solving and streamlining solutions to the organization’s Human Resources department through big data analysis. The works of this team includes Project Oxygen (identification of the best leaders) and Project Aristotle (for the composition of the best work teams). The research pertaining to these studies is available on a dedicated website and has led to similar departments being established in other major companies like Amazon, Facebook, and Microsoft among others.

So, in essence, Artificial Intelligence has stepped out from the shadows of being research-oriented to be a practical driver of change. This is mainly due to the rapid evolution of technology pertaining to machine learning, big data, and neural networks in recent years thanks mainly to the rapid evolution in recent years. Major tech firms, like Apple, Google, and Amazon, are putting a lot of money in developing AI solutions and the results are beginning to emerge in the form of general-use solutions like virtual assistants (Sir, Google Assistant, Alexa), voice and facial recognition, as well as real-time translation. To put this into perspective, Sundar Pichai, Google’s present CEO, has recently said that AI will be more important to humanity than electricity or fire.

Furthermore, according to data shared by Glassdoor, an employee-oriented network focused on anonymous information sharing, data scientists have held the most valuable jobs in the US for the third straight year. The demand is high in the industry with an average of 4500 openings available to data scientist at any given time and a median starting salary of $110,000. And this demand is materializing into machine learning solutions being developed in several industries. Human Resources is one of them as companies like Glint, PhenomPeople and Peoplise provide AI solutions for selection, talent management and improvement of employee experience.

In conclusion, we are still decades away from an AI solution that can completely emulate the workings of the human brane, but in the field of information analysis at least, it is possible to write algorithms and create neural networks that can spot connections in data that a human analyst may miss. These solutions are particularly powerful when applied to unstructured information, as is the case with the data and information available in the area of Human Resources. That said, such solutions are burdened by their own forms of prejudices induced by data bias, so it is still optimal to form a balanced human-machine collaboration where the prejudices of either are masked as well as possible.