[avatar user=”Prasadh M S” size=”thumbnail” align=”left” /] Prasadh M S |

Can AI Hire for ‘Potential’?

Cricket in India, thanks to its colonial shadows, for long remained a cosmopolitan game. Though players from small towns eventually found their spots in the team, the Captain’s cap remained elusive for a while longer. Then came the small town star from Jharkhand, fierce yet humble. Getting picked for Captaincy in his 4th year with the team, toppled all analytical models and data driven predictions. A nation in surprise and lobbyists in shock, couldn’t have guessed the potential that he went on to show. Call him “Mahi” or “Thala” or “Dhoni”, he would remain one among the most decorated Captains Indian cricket has seen. If not for the potential that some selector saw, and if data science had its away, we might have had to wait a little longer for blockbusters from Dhoniwood!

Haven’t you, at some time, either been hired for or hired others for potential?

Future of Work

Future of Jobs 2022 study from Ernst & Young reports 46% of India’s skilled workforce will be deployed in new jobs or jobs with radically changed skillsets That’s a whopping 270 Million who would be in jobs drastically different from their past profiles. Without qualification or previous experience as direct markers, hiring will continue to be for ‘potential’ rather than profile fitment.

A look at the list of high paying jobs today that didn’t exist 10 years ago, would show that hiring for potential is not a new trend. Top end Digital Marketing Specialists, Big Data Architects and User Interface/User Experience Designers were all hired for potential rather than for prior experience. That poses a question and a challenge for data dependent technologies like Artificial Intelligence (AI) that are re-defining recruitment – Can they spot ‘potential’?

Candidates are CVs?

The candidate CV continues to evolve into multiple forms and formats. However, the CV’s exact role in recruitment still remains to be deciphered. How often, after initial screening, do we go back to the CV? It’s popularly stated that the experienced recruiter spends a little over 6 seconds scanning a CV before shortlisting or rejecting. In face-to-face interviews the CV is all but a cheat sheet for the interviewer to pick up questions.

A study by the American Sociological Association found that a candidate’s similarity to the interviewer and the organisation affects final hiring decisions. Recruiters and Hiring Managers consciously look for shared tastes, experiences, leisure pursuits, and self-presentation styles as well. Needless to say, none of that resides in the CV and the decision for or against is not entirely data-driven. After all, Candidates exist outside their CV and we hire the person and not the CV.

Enter Artificial Intelligence

The cliched debate about recruitment as science or art is more relevant now than ever before. As we talk, CVs have become infographics, social media defines the candidate persona, blockchain career verification is arriving and career net-scores are getting computed. Access to enterprise data on past hiring processes is easier than before. It is this access to machine readable data that inspired the Artificial Intelligence (AI) wave in Recruitment.

With such elaborate data trails around candidates, shouldn’t hiring be just about rating, ranking and skimming the cream from the top? After all, the who-why-when-and-what of hiring decisions is all accessible for AI to learn from and mimic. We seem to have a strong case in favor of data science. Or did the science vs art debate just get complicated?

Monkey See Monkey Do

AI’s core plot is to get machine learning (ML) to continually crunch past and current hiring data and pickup patterns behind hiring decisions. Once the patterns are known, AI should be able to fluently mimic them on a different data set. For example…how often do you struggle with extracting first names from a full name column on a spreadsheet? Excel’s in-built AI today does it for you after you’ve manually done the first 2 rows. Now that’s a poor example, to illustrate what AI has promised to do for Recruitment. AI is expected to replace 16% of HR jobs within the next 10 years, says a popularly quoted report from Undercover Recruiter.

Does AI’s ability to read the what and how of a hiring process, imply it can autonomously do it on its own. I can hear the AI Evangelists scream, when did AI ever claim to make hiring decisions and replace recruiters? Well… if that’s not the aspiration then am afraid AI would already be redundant in this space. Lets not forget that the HR space has one of the largest shelf of process automation suites and data processing tools. AI cannot just harp on replacing ‘mundane’ HR tasks any more.

The ‘Potential’ Challenge

Gauging candidate potential is a qualitative (a.k.a subjective) process and it cannot be read without bringing the Recruiter’s persona into the equation. The evangelists say that’s exactly the bias that AI will eliminate. Well… this bias is good bias and am not sure if we would like to trade it off.

There are listicles published every other day about Ways AI Will Change Recruitment Practices. So let me not use real-estate debating them here. However, the observation is that most are in denial of the big fat fact about Candidate Potential being a key driver for hiring decisions. ML and AI do not catch or gauge candidate potential…not because they can’t but because they don’t have the means to learn it from. Gauging potential is contextual to the organisation, its culture, current stage of growth, co-workers & manager fitment and so on. Attempting to convert these to machine readable data would be a wild goose chase for AI ventures.

Artificial ‘Intuition’?

Experienced recruiters and hiring managers pride over the instinct and intuition that they develop over time. Instinctively picking a candidate with fitment and the intuition that drives a gut feel, are beyond the realms of data science. The ability to visualize a candidate’s potential as part of the organization and assess the resonance or dissonance is not ‘machine learnable’. AI’s gauging of potential and visualisations of such fitment, if attempted, can at best be Artificial.

AI is evolving fast and we look forward to what else it can do. However we also remain cautious about positioning it as a magic potion for every perceived ailment in the HR and Recruitment process.

So…do you think #AI would have suggested Dhoni as captain?