Job interviews by AI are a reality—could they turn into a preference for job seekers too?

The Mumbai-based accountant has been interviewed 27 times in her career—25 times by humans and twice by artificial intelligence (AI). She was initially nervous, and in her first AI interview, the American-accented artificial voice asked such a multi-pronged and convoluted question that she was overwhelmed. She logged out and never attempted that interview again, despite reminder emails from the firm behind the exercise.

It is unlikely that she would have logged out or done the closest equivalent had there been people on the other side; that would have been unprofessional and rude. But mercifully, AI bots do not have feelings… even if they sometimes seem to.

It was the chartered accountant’s second interview, this one for a quality and audit role, that began to change her mind. As she recounts the experience, staring at a screen-saver while an American-accented voice asked questions felt oddly comforting.

Why? I asked. Didn’t she miss gauging how the interview was going from the expressions of panelists?

No. The AI interviewer did not have a face, its tone didn’t give away how the interview was going and there were no smiles or sighs to interpret. In fact, after every answer, the AI bot summarized what she had said. “In a few questions, I fumbled. But the way the bot summarized my answers, it sounded right. It seemed like I had given the correct answer even if I hadn’t. No interviewer would do that."

As a business journalist, I have tracked Human Resources management for about a decade and written about how interviews have evolved—from the classic “Where do you see yourself five years hence?" to ethical dilemmas like “Who would you rather save: a senior citizen tied to a railway track or a group of men tied together on a different track if you can save only one side?"

But the concept of being interviewed by AI seems very complex. More than that, I’m not sure I would trust myself with my answers when I know there are no second chances, or that the bot has a fixed menu of questions and no digression would be allowed. I can’t read its expression or get cues, yet the algorithm would be picking up mine.

The process feels daunting—even tougher than psychometric testing rounds or facing psychologists on a human interview panel.

And yes, there will probably be human managers reviewing the bot’s recordings and assessments before finalizing decisions.

But that does not explain what prompts recruiters to hand over the reins of interviewing (their most crucial selection tool) to an algorithm. Is it a quest for perfection in finding the right fit for a role? Or is it another way to pass on recruitment risks?

Are recruiters giving up the onus of selecting a candidate who simply feels right for the role on offer? This is not always about answers to questions; often, it’s about sensing that the candidate may or may not click, but is likely all the same to bring in fresh ways of thinking. This is a finer judgement call that AI may be unable to grasp.

Under an overload of applications, perhaps AI does a useful job of basic screening. AI also promises significant time and cost savings by replacing human interviewers, especially for junior roles where primary competence matters more than other attributes. And while India Inc insists that middle- and senior-level profiles, which require collaborative and leadership qualities, will still be interviewed by humans, we can’t say where this is headed.

A senior partner at one of the top three global headhunting firms told me that as AI gets more sophisticated, it will analyse candidates better than human bosses can, be it “detailed personality assessments, predictive insights into opportunities and risks if the firm hires the candidate" or “the open-ended questions that allow deeper checks and balances." He admitted that he fears a critical part of his company’s role in recruitment may vanish.

At a time when companies take six to eight months on a CXO hire, some of them may prefer to let AI take the call. It’s easier to blame AI if the candidate doesn’t work out than to question the managers who made the decision.

Eventually, the big advantage of AI recruitment could turn out to be that it is less biased—a challenge deeper than India Inc openly admits.

In Hrishikesh Mukherjee’s film Gol Maal, the protagonist dons a fake moustache, avoids Western attire, hides his love for football and even pretends to have a twin brother, all to land a coveted job by aligning himself with what he knows the decision-maker prefers. The film, something of a Lemony Snicket-style chain of unfortunate events, is a laugh riot, but it mirrors the deep-seated biases interviewers often carry.

Unless AI is trained on biased feeds, it could rise above all that.

The author writes on workplaces and education at Mint.