The Limits of AI in Hiring Why Human Judgment Still Matters

Abhishek Nag is not a conventional HR leader. An engineer turned entrepreneur turned HR executive, he currently serves as Global Head of HR at Fornax, a multi-entity services organization with more than 40,000 employees across 16 countries. That cross-functional background gives him a systems-level view of how technology, talent, and decision-making intersect at scale.

In a recent episode of the Talented Podcast by Glider AI, Abhishek joined host Joseph Cole to discuss how artificial intelligence is reshaping recruitment, skills validation, and workforce trust. Rather than leaning into hype, his perspective was grounded, pragmatic, and occasionally provocative. AI, he argues, is neither a silver bullet nor a threat on its own. The real risk lies in how leaders choose to use it.

Below are the key insights from the conversation.

AI Will Not Replace Recruiters. Recruiters Who Use AI Will.

Much of the anxiety around AI in hiring centers on job displacement. Abhishek reframes the issue more bluntly. The real danger is stagnation.

“AI doesn’t take away your job. Someone using AI does. If you stagnate and don’t use AI to augment your skill set, you’ll become obsolete sooner.”

Rather than eliminating recruiters, AI is absorbing repetitive, time-intensive work such as scheduling, basic screening, transcription, and data processing. This shift frees recruiters to focus on judgment-driven work that technology still struggles to replicate.

Gartner supports this view, noting that AI is increasingly being deployed as a decision support layer rather than a decision-maker in HR, augmenting human work instead of replacing it.

For talent leaders, the takeaway is clear. AI fluency is quickly becoming a baseline expectation, not a differentiator.

Screening and Scale Belong to AI. Judgment Still Belongs to Humans.

Abhishek draws a clear boundary between where AI adds value and where it should stop.

“Certain skills that are easily analyzable through technology can be delegated. Behavioral attributes and complex, ambiguous roles still require human judgment.”

AI performs best in structured environments where job requirements are clear and measurable. It can process high volumes of applications, detect patterns, and surface outliers that humans might overlook. But non-linear career paths, cultural nuance, and contextual decision-making still demand human involvement.

Deloitte research echoes this balance, showing that organizations achieve better outcomes when AI enhances human decision-making rather than replaces it.

The future hiring model is not fully automated. It is selectively automated, with humans stepping in where context matters most.

Bias Exists on Both Sides. AI Can Help Expose It.

Bias is often cited as a reason to resist AI in hiring. Abhishek challenges that framing by pointing out that bias already exists, with or without technology.

“There are two kinds of AI bias. One is programming bias. The other is training bias. But humans also carry bias into every conversation.”

AI systems can inherit bias from training data or design choices. At the same time, they can act as mirrors, flagging inconsistencies or patterns that reveal human bias during interviews and evaluations.

McKinsey notes that AI can help reduce bias when systems are transparently designed, continuously audited, and paired with human oversight.

The responsibility lies with organizations to understand how their AI is trained, what data it uses, and where human intervention is required.

Candidate Fraud Is Rising. Verification Must Scale With It.

AI has not only changed how companies hire. It has also changed how candidates attempt to game the system.

“Earlier, there were smart cheaters and not-so-smart cheaters. Now everyone has become a smart cheater.”

From impersonation to real-time AI assistance during assessments, fraud has become harder to detect manually. Abhishek emphasizes that AI-powered proctoring, behavioral signals, and pattern detection are now essential safeguards.

However, detection alone is not enough.

“An AI can help answer a question. It cannot explain it.”

The most effective defense combines AI monitoring with human follow-up, asking candidates to explain their reasoning, walk through decisions, or defend outcomes. This approach separates surface-level output from real understanding.

HR Must Lead AI Adoption. Change Management Is the Job.

Enterprise AI adoption is not just a technology rollout. It is an organizational change exercise.

“There’s a strong case for HR to lead AI transformation because there’s a lot of change management involved. People fear job loss. Psychological safety matters.”

HR leaders sit at the intersection of culture, compliance, and workforce trust. That position uniquely equips them to guide responsible AI adoption, address employee concerns, and define ethical boundaries.

PwC research shows that trust and transparency are among the biggest factors influencing employee acceptance of AI-driven change.

Leadership ownership, not just technical capability, determines whether AI initiatives succeed or stall.

AI Will Become a Hygiene Skill. HR’s Role Will Elevate.

Looking ahead, Abhishek believes AI proficiency will soon be assumed rather than highlighted.

“AI will become a hygiene factor, like email or Zoom. Everyone will use it to a certain degree.”

As internal HR tasks are increasingly automated, the function itself has an opportunity to shift outward. Less time on internal administration. More time on employee-facing work, quality conversations, and strategic decision support.

This shift aligns with broader workforce trends showing HR moving from operational execution to strategic enablement as technology absorbs administrative load.

What Leaders Should Take Away

Abhishek’s message is not that AI will solve hiring. It is that hiring without AI will not scale.

The organizations that succeed will not be those that automate everything, but those that understand where automation stops and judgment begins. Leaders must invest in AI literacy, continuously audit their systems, and treat trust and verification as complementary rather than opposing forces.

“Trust, but verify.”

That principle may define the next era of recruitment. Abhishek has laid out the challenge clearly. What comes next is how leaders choose to act on it.

Stay tuned for upcoming episodes of the Talented Podcast, where more HR and talent leaders share how they are putting AI into action across hiring and workforce strategy.

Interview Transcript

Joseph Cole (00:01)
Hello Abhisek, ⁓ thanks for joining us today. Could you introduce yourself, ⁓ your professional background, where you work, what you do, and then maybe something that, some personal info that people can find in Google.

Abhisek Nag (00:15)
Thank you. Thanks for having me here. the pleasure is mine. Quick introduction. My name is Abhishek Nag, as you can see, and I currently lead the HR function globally for our services organization called Fornax. We are a group company with seven different organizations across different service sectors with 40,000 plus employees globally and in 16 different locations and we growing.

⁓ Something that you won’t find about me normally is that I’m an engineer turned entrepreneur turned HR so have had a varied experience in different industries right from an education to a healthcare to a technology both product and services and enjoying my journey all through.

So that’s about me and beyond my work hours, I love to volunteer for lot of different social causes. So that’s a little bit about me.

Joseph Cole (01:22)
Awesome. A few questions there. Interesting pivot from engineering to HR. Is that pretty common in India or because it’s definitely not common in the US?

Abhisek Nag (01:36)
⁓ There are quite a few examples like that, ⁓ not too common, but we have ⁓ a sizable population making that change over a period.

Joseph Cole (01:51)
Yeah,

interesting. And then what about the causes that you support? Anything that you want to share about those?

Abhisek Nag (01:58)
primarily

on education. So, yeah, education for the underprivileged.

Joseph Cole (02:01)
Very nice.

Great, okay, great. So let’s jump into the first question here. ⁓ From your perspective, ⁓ why is AI becoming such a hot topic specifically in recruitment and HR and talent management today?

Abhisek Nag (02:18)
So if you look at the genesis of AI, there was a time when artificial intelligence came into being and we started using it for predictive, for perspective information. Gradually with the advent of Gen AI today, it has started augmenting a lot of intelligence that otherwise would have taken ⁓

time and bandwidth.

for the recruiters, for the HR folks, on things which could be delegated to a technology. Now, it started with the basic automation, but it has, over the years, taken up the role of more strategic functioning. And off late, have seen a lot of prevalence of AI tools to support some of the functions which otherwise were thought of ⁓ not to be possible

Joseph Cole (03:06)
you

Abhisek Nag (03:19)
For example, a lot of good recruitment screening happening using AI.

The previous version of technology was that one size fits all technology and with that Gen AI, the magic is that it’s extremely customizable and trainable based on the organizational context and the needs and which is where a great use case of Gen AI, comes in play. ⁓ If trained well with the data set of the particular organization, it starts acting favorably like more like a human.

to screen, to filter, to make certain choices as the organization would do.

Joseph Cole (04:08)
Yeah, you know, kind of stepping back a little bit and you you shared a little bit about, you know, you moving from engineering to HR and I think this is somewhat related to, you know, what we’re talking about now with AI and GenAI becoming, you know, so much a big part. So I was taking away a lot of the entry level positions, right? Even in HR and recruiting, do you have any advice for people trying to get into HR or into recruiting?

Abhisek Nag (04:35)
I think it’s quite cliched now that AI, only AI doesn’t take away your jobs, it’s somebody using AI does. So essentially ⁓ the message is very clear loud that you start using AI or else if you only stagnate being just doing the part of it and not using an AI technology to make, to augment your skill set, then probably you’re going to be obsolete sooner.

Joseph Cole (04:46)
Yeah.

Abhisek Nag (05:05)
So an AI tool is to enable a human recruiter faster, better, and making more practical pragmatic decisions. And I think that’s going to be the future.

Joseph Cole (05:18)
Yeah, and this is off script here, if you can’t answer, obviously we’re good. ⁓ What AI tools do you leverage as an HR leader, and how do you leverage those tools as part of your job?

Abhisek Nag (05:34)
So there are quite a few that I would use.

⁓ many different functions where there are AI tools that are used right from recruitment to learning and development to ⁓ in general basic HR data processing, analytics. So every function of HR today, even how employee queries, self-services, et cetera, every function of HR is being impacted by AI largely except, and in all functions,

decision making is still human. The intelligence comes from the AI helping whoever is making the decision, helping make the decision better, faster and ⁓ probably more pragmatic.

Joseph Cole (06:29)
Got it. Do you have any favorite prompts to use?

Abhisek Nag (06:36)
So I mean, I personally love GROK for all my set of tasks, but in general, all my team members use Canva for designing, for example, we use a lot of other tools for.

Joseph Cole (06:49)
Okay.

Abhisek Nag (06:55)
for whether recruitment or indeed we have different tools. I don’t want to specifically call some of them out, but we have multiple ones. But I think irrespective of whichever function we have either copilot, have a corporate copilot. So copilot is again something that we use frequently, multiple other tools. again, depending on what the use case is. And I’ve realized now that depending on the need and use case, are a variety of tools.

Joseph Cole (07:01)
Yeah, yeah.

Abhisek Nag (07:24)
presentation probably come out be a good one to go for or things like that even I’ve seen some the recruitment products like light and all they’re also great ⁓

Joseph Cole (07:25)
you

Awesome. Well, thanks. I know that was a little off script, but just more curious as we were talking. the next here is AI’s role in recruitment. So how do you see AI changing the way companies attract and hire talent over the next five years with AI?

Abhisek Nag (07:57)
So I’m going to be slightly ⁓ provocative here that it’s going to be a tough time for companies with using AI. And the reason is ⁓

When you put up a system to evaluate screen, there’s going to be another system coming up and engineers are quite ⁓ known for doing that. There’ll be another system to game that. So essentially that’s going to be a system versus system play. How do we beat that is fundamentally applying human judgment in the process where needed. And that’s going to be the critical aspect.

Joseph Cole (08:29)
Yeah.

⁓

Abhisek Nag (08:40)
There are going to be great use case of AI on ⁓ trading the data set with the organizational context and all which will help ⁓ look at certain ⁓ outlier, which will look at certain. ⁓

linear carriers which are aligned to the organizational structure, but there will be 10 % 15 % nonlinear carriers and it will require a human decision making on those judgment calls. But I think the basic functions of scheduling and even

Making sure that the there are there are edge cases being covered those those cases AI would be a great tool for example transcribing an interview people forget to give helping them but creating that set of things or even

using ⁓ prompts for interviewers, etc. Those are fantastic features that AI would help keep track of and help both the interviewee and the interviewer to be on the best of their game. And I’ve seen that happening in different tools where

the interview and the interviewer gets prompted on segments on how to frame it. there’s today with the geographical barrier being upset, there is someone from a different geography interviewing another person and the dialect may come in play, but with the use of AI you can ⁓ kind of translate the question or have it live screening which may make it easier for the interviewee to understand it, respond better.

So there are a lot of good use cases of AI which can help the experience for both the interviewer and the candidate better and be the best of their versions to be finally evaluated.

Joseph Cole (10:40)
Yeah, and you answered some of these with the transcription and the translation aspect, are there any other, I guess, during the recruitment process, there’s so many different aspects of it, like the interviewing, the screening, the skill assessment, the fit assessment, behavioral assessments, right? And then there’s just the automation. Do you have any perspective on what part of the recruiting to hiring process is best suited for AI or automation versus what

is better suited for human judgment.

Abhisek Nag (11:13)
So skills, certain skills which are easily ⁓ analyzable through technology, think those things can be delegated. There are behavioral attributes which probably need to be validated by human. Also, helps, AI helps in detection of fraud in a great way. Today, ⁓ we have hard cases of impersonation, we have hard cases of malafied intent, et cetera, and all of that can be ⁓

protected

using AI, easy to catch ⁓ beyond human eyes or things. And back to your question, what kind of roles? I think there are certain roles which are very structured or indeed have specific job sets. In those cases, you can easily create assessments which can be treated through AI. The roles which itself has a lot of

ambiguity in the nature of the role or complexity involved those kind of roles would require human judgment more than automation.

Joseph Cole (12:20)
Right. Do you think there’s a, I mean, there’s definitely a balance between leveraging AI versus human. I guess from your perspective, you know, I feel like humans are inherently biased, right? We just naturally are just, you know, how our brains function, right? But there’s a lot of talk about like AI being biased and yeah, it makes sense. You know, mentioned some of this, like the data needs to be good. Do you have any perspective on like human bias versus AI bias and you know, like

Abhisek Nag (12:45)
So there are two kinds of AI bias. One is a programming bias and one is a training bias. So there’s this.

There’s certain boxes that while it was programmed, it will automatically reject. And there is then the bias of training data set that comes in. And depending on what the AI, and generally an AI is mostly trained on linear carriers, for example. So the moment you get a person who has a nonlinear carrier path and things, may not consider it even. So it’s important for.

Joseph Cole (13:09)
Hmm.

Abhisek Nag (13:21)
a human too.

analyze that data set continuously train that and I believe with the the advent of ⁓ newer models of AI the system will evolve and continually get better so it’s a responsibility for us to train it with the right data set with the prompts and make it better. Coming to the human bias side I think AI can help catch the human biases show it to to a human say that these are it’s kind of a mirror reflection showing it that these

Joseph Cole (13:47)
Yeah.

Abhisek Nag (13:54)
These

are the things that you might have overlooked or you might have ⁓ considered much and then make a judgment call. So essentially today when you are having a conversation with the candidate, every human being having a conversation will have certain biases that play in. Now we do not know what all biases probably have gone in that conversation and AI can show it up and help identify that. Then it becomes again becomes a judgment call for the human

to make that decision ⁓ and that still will remain human in my opinion.

Joseph Cole (14:29)
Yeah, and then obviously you as an HR leader, and I think this really does build the case for HR and TA to be involved in AI implementation across the enterprise, is because you have that human aspect, right? You’ve always have pulse on company culture and all these other compliance things. I guess then with that, ⁓ what advice would you give to other CHROs or HR and TA leaders, you know, when they’re looking at AI and like considering it and concerned about

you know, some of the potential issues like bias and AI.

Abhisek Nag (15:05)
So I think ⁓ one aspect is look at what training data set is being used to train your particular models because that becomes critical for you to see whether

whether the AI is suitable for you. mean, for example, if you’re hiring for indigenous or dialects or local languages which are inherently not, who are non-native English speakers, for example, is the AI trained enough with enough data set to assess that kind of profiles or things? otherwise, even though they may have a dialect or may have a heavy influence of certain things,

Joseph Cole (15:42)
Okay.

Abhisek Nag (15:49)
not find them good candidates. So essentially look at, as each us look at, what is the data set that the AI has been trained on? B, what is the use case that you are going to use? And if you’re not sure, start with something low-hanging which does not impact.

in the decision making rung has a lower impact on the decisions that you’re making and gradually go higher up in the scale. So start ⁓ small and use it, get comfortable and then take it from there.

Joseph Cole (16:16)
Yeah.

Right.

You know, this is something that we’re noticing a lot of, ⁓ know, one of my friend and also professional colleague, he runs an analyst firm. ⁓ They have an AI council and they’re really making the case for HR to really lead AI transformation in the enterprise. ⁓ You know, how do you, what, mean, is AI, is HR at your company involved in, you know, enterprise wide AI implementation? And then like, how does ⁓

target a seat at the table.

Abhisek Nag (16:55)
It’s a good question and a great one that whether the seat at the table. I think it depends on the interest of the individual and any business leader could be an implementation leader. There is always a strong ⁓ case for HR to be that person because there’s a lot of change management involved which is where the psychological safety comes in play as you asked about earlier that people fearing lost jobs and all. So that fears

is always at the back of the mind for people when an AI implementation is happening. So essentially an HR leader has ⁓ ability to coach people and manage that change better. But having said that, it fundamentally depends on whether as a leader, do you really believe on AI and are you going

Joseph Cole (17:34)
Hello? ⁓

Abhisek Nag (17:49)
take charge of that decision and make it happen because you believe that it’s going to be for the betterment of the organization. So that belief has to be there to get that seat on the table. once you have that belief and once you’re an ⁓ advocate of that, you’ll probably be the first proponent to make it happen and drive it through the organization.

Joseph Cole (18:09)
Right, not a good,

yeah, really good perspective, thanks for that. Now getting back to the scripts, I appreciate the detour there. So do you think AI, as Jason always said, will ever replace the recruiting function, or at least what aspects of the recruiting function do you think AI will replace within the next year, basically? It’s moving so fast, right?

Abhisek Nag (18:31)
So it will take away the.

the basic tasks of scheduling, basic tasks of doing a perfunctory call, will require value addition that, okay, this is what the real data I have and also targeting passive candidates. The part of selling an awkward candidate, that’s still gonna be human. The assessment may still be more AI driven in the future, but the part of recruitment where you actually go

Joseph Cole (18:55)
Mm.

Abhisek Nag (19:05)
and ⁓ sell the role and make the candidate excited about the role, that will still be more human. Then the final decision of selection will be human. But intermediaries steps may get delegated to an AI faster.

Joseph Cole (19:21)
right makes sense to me as well

Okay, so let’s jump into the next area. Skills validation and candidate assessments. ⁓ So many candidates are using AI tools to enhance their resumes or even fake their skills, right? How do you think AI can be used as a tool to then truly validate their skills, right? Based on all this other pretend stuff, right? And it’s so easy. Like for example, ⁓ we’ve had customers where once upon a time, you know, they’ll get like 200 resumes over the course of two

months. Now they get thousands of resumes in two days and they just have to close the job posting. So I guess like how can AI be used to like circumvent all of this deluge, deluge, I can’t even speak English right now of like, you know, candidates who aren’t qualified.

Abhisek Nag (20:12)
So I think there are a of cases here. One is that there’s certain series.

There are certain series right now where there’s one segment of people who use AI to enhance their service and there are certain who don’t. And then the next step is people who, if you put up an assessment, there’ll be certain people who use AI to complete that. There are certain people who don’t. Now, fundamentally, people who use AI, think the mission should be favorable for them because at least they are more attuned to it. They’re more… ⁓

Joseph Cole (20:42)
Bye.

Abhisek Nag (20:50)
future thinking, future forward. I don’t ⁓ hold it against people who are using AI to solve problems because they’re going to use AI in the workplace also to do certain things. Now, the question is, can you, even if you use a vibe coding tool to code, can you really understand the code? And the beauty lies in you can solve a case study, but can you explain the case study? So that is where the…

Joseph Cole (20:52)
you

Abhisek Nag (21:16)
lies and I think that assessment can be game changer that you can can give a case you can give an assessment and let people do whatever tool that they want to use come and explain that and that that kind of differentiates between the real and the fake ones

Joseph Cole (21:20)
No.

Yeah,

makes a lot of sense to me. It’s truly that critical thinking, the aptitude, the cognitive ability, or the soft skills that are becoming more more important. Then ⁓ if people are using these AI tools, how do you assess, you know, from your standpoint, the critical thinking or the cognitive ability or the things that are less tangible, like, you know, the EQ?

Abhisek Nag (22:00)
So.

Again, depending on the roles, there are different ways. But one of the great things that I’ve seen in this, and we’re trying to kind of emulate that also, is do a dry run of the work. So essentially give them a real work, real scenario, ask them to work in the real setup and see how they perform and do. And that’s the best case because it’s more close to the reality. I think there are lot of great organizations who do that.

⁓ And more so over the time, with time I think, at least for the early career roles I see there’s a lot of change going to happen in the future where internships or shadow work will become the route to get people. For the experienced people I think the reference checks will become more critical, knowing their past work and examples of that will become critical and things like that.

Joseph Cole (23:00)
Yeah, it makes sense to me. So jumping into fraud, you mentioned a little bit of it, right? And we kind of alluded to some of this with the use of AI and like, what’s the balance of like using AI and critical thinking to do a better job? Have you seen or witnessed any candidate fraud because of AI? Has it increased? Has it become like really, you know, interesting where you’re like, wow, is that a real person or do have any examples?

Abhisek Nag (23:28)
So I think…

What has aggravated is I mean impersonation or fake candidates fake things were always there even before AI but what has ⁓ what has kind of become difficult is to catch the people because there are earlier days they would have to kind of have someone helping them to answer now with the AI you kind of ask a question there is a device beside you which is kind of answering or

saying and it’s difficult to it’s become more difficult to catch but I think again it’s it’s kind of a game of catch-up on one side that the interviews training tools are getting better to catch those kind of cheatings similarly people are becoming ingenuous to add more ways to cheat so it’s kind of constant catch-up in that way but use of AI has

Joseph Cole (24:00)
Yeah.

Abhisek Nag (24:28)
What has happened is people, they were,

kind of I would say ⁓ people who are smart cheaters and people who are not so smart cheaters, that differentiation has reduced. So everybody has become a smart cheater now and therefore it becomes difficult to catch that. But you, as I said earlier, right, you get, when you kind of question the output of the work and ask them to explain, that’s where ⁓ it gets tested really well and I think.

Joseph Cole (24:41)
Yeah.

Yeah.

Abhisek Nag (25:00)
An AI can help you answer the question but not explain it in the way that when you go deep into it.

Joseph Cole (25:08)
Yeah, no, no, interesting. We’ve seen a lot of it, obviously, with our customers, you know, and they leverage our AI proctoring tool. ⁓ I guess what are some telltale signs that you’ve seen or you’ve, you know, heard of, you know, with your with your team, right, to find fraud or like see it actually happening?

Abhisek Nag (25:31)
I mean, you get all sorts of right from documentation fraud to impersonation or someone trying to use it in the runtime to… ⁓

answer questions which they don’t even know of. So all kinds of things come up and some are easy to check, some are difficult they said. ⁓ It’s always and sometimes it becomes you get to kind of

go through the process and then realize it at a later point. You may not realize it at the right, the get go. It’s happened and it’s always a loss of time, effort, energy, and cost. So ⁓ there’s also a bit of things that come up that.

Joseph Cole (26:16)
Yeah.

Abhisek Nag (26:23)
Sometimes we question that is it ⁓ like ⁓ is it the right ⁓ thing that people are doing it and and things so ethical questions come up that how do we how do we ensure that ⁓ we get that because We get the right set of people who? Who are not? ⁓ Kind of on the borderline of ethics. I mean, where do you draw the line like is it? ⁓

somebody is using one question or take a help versus somebody like completely impersonating. So obviously there are some which are clear cases, then some there might be one smaller thing they have issued and you have a business. So we try and make sure that we are an ethical organization and we do not consider any, we take every serious violation of these seriously. And that creates a challenge that you have hired someone or you have gone through the process and then you realize,

have to redo the process again with the entire flow. So it’s a hard… I guess if there was something that can either deter people to do this or build more trust within the organization to be able to kind of cut down on that, that would be really good.

Joseph Cole (27:46)
Right. So, you know, obviously there’s a lot of this and companies are obviously needing to be very cautious about the people that they bring in, ⁓ you know, with AI or, you know, without AI, that’s, you know, obviously the most important part of the company. ⁓

But then, you know, obviously there’s the recruiting process. So how do you balance all of this like extra steps of like ID verification or AI proctoring while ensuring a smooth candidate experience?

Abhisek Nag (28:20)
So, ⁓

When you say there’s a balance, the balance is about that we give you freedom to operate, but that does not take away the responsibility of you as an individual. For example, do you genuinely need to, in an interview, I’m taking it slightly in a hypothetical example here.

Joseph Cole (28:38)
Yeah.

Abhisek Nag (28:52)
you’re doing a proctored and somebody says I want to really need to take a break and step out. Do you consider that as a…

and they can go and seek help and come back or they are actually genuinely in for the point is that you consider a benefit of doubt that yes the person is probably genuinely in for but it comes from the fact that are there generally any such any issues of these sort of impersonations or fake things there are multiple signs of it it’s not just one thing and

An AI tool, while the human eye may miss one of those, the AI tool can flag some of these things and prompt the human to ask appropriate question. And I think that’s where the balance comes in that trust but verify. like if you are small, it shows me a sign that, hey, there is something wrong going on. would.

Joseph Cole (29:42)
Yeah.

Abhisek Nag (29:48)
be overtly cautious, ask questions, and then if I see something, then I may want to kind of be more conservative on the freedom that I’ve given. Or if on the other side, there’s no red flags, there’s something, then I’m more lenient in the process. So I think AI helps to bring that balance that should I, how much freedom should I give? And I’m talking about a proctored interview as a use case. There can be multiple things. So essentially,

Leave room for certain freedom but have the system flag you and then take a judgement call whether this is a right, genuine or a fake thing.

Joseph Cole (30:30)
Yeah, I like how you put a trust but verify, right? Especially when there’s questionable things. I think also part of that is providing the communication, the transparency upfront. So they’re not like surprised, but like, why are you doing this? Is that something that you do as well or in your own recruiting process?

Abhisek Nag (30:42)
Yeah. Yeah.

Yeah, we build in transparency. think transparency is something that’s built in from the get go. Everything that we put across is laid out transparent, that this is the rule, this is what is expected, there’s some rules which you use, some rules which don’t. So all of that is clearly called out so that the person knows what are the boundaries and what is expected, what is not, and what part of things are non-negotiables for us, what part of things are open.

And that helps to even take decisions. And we have tried to now, trying to put up a process where we can provide feedback to candidates. And as a part of that, we’d want to give feedback to the people. If any of these non-negotiables are not met, where ⁓ did it ⁓ miss out and things like that, so that it’s transferred to the candidates.

Joseph Cole (31:24)
and

Yeah, I

like that. ⁓ Specifically the non-negotiable, guess. Can we unpack that a little? Do you have an example of what you might consider a non-negotiable at your company?

Abhisek Nag (31:56)
⁓

So for certain roles we have, and being a central cognition, we have customer data protection as a very strong case. So any sort of, ⁓ and then in those kind of roles, how you use data, what you do with that, and things are non-negotiable. if there’s anything, if you do not understand how data privacy works and stuff like that, that becomes a non-negotiable for certain roles. So there are multiple

in each role there are different conditions which are non-negotiable. So for example, even if I’m hiring somebody in HR as a competent, then how they treat data is going to be sacrosanct for me. Again, ⁓ there are roles where, and as a value, we value respect a lot, ⁓ the value, so anybody who is disrespectful in their behavior attributes becomes, ⁓ that’s a non-negotiable for us. It doesn’t matter which role, what hierarchy at all. ⁓

Joseph Cole (32:47)
Yeah.

Yeah.

Abhisek Nag (32:58)
mix of both soft skills. There are certain roles where ⁓ attention to detail is very critical. So even a minor miss is a miss and cannot be considered. So there are certain ⁓ technical or skill-based aspects. There are certain functional or behavioral attributes, both that are non-negotiables depending on the role.

Joseph Cole (33:22)
Yeah, I’ve never heard it from that standpoint before, but I think it makes a lot of sense. And it’s also something that you can tie back to like company values and core values, right? And then it gives you like a nice strong basis or foundation for like how you hire and who you hire. So that’s a really nice way, I think. ⁓ So one question here about just general risks and challenges, like what major risks do you see about AI implementation in the human part of business, like, you know, recruiting, tele-management?

egg car.

Abhisek Nag (33:56)
So one the major challenges people not understanding where to use AI and thinking that it is going to solve everything and start using it for everything without understanding the impact of it. So as simple as that, if you ask AI to, and this was in the media recently, there’s some article ghostwritten by someone and. ⁓

Joseph Cole (34:11)
Yeah.

Abhisek Nag (34:24)
At the end, say, if you ask me to write more, I can do that. So if you do a specific example, ⁓ there are systems being implemented, whether it’s an LND or recruitment or in other cases.

how do you handle when the AI is in doubt? So if you understand a system, you’d configure it to handle that, make sure that it is ⁓ straight with right data. There are how you input the information, what are you asking it to do, what are you expecting it to do, you know where can it falter. So all this knowledge helps you to use a system better. And the flip side is if someone who do not know it or is not

aware of it, we’ll go and use it and then ⁓ when it backfires it has an impact on people, has an impact on lives and on the organization, the business and ⁓ you probably do not even know where that has gone wrong and that becomes very detrimental and challenging then.

Joseph Cole (35:27)
Yeah, and actually I do have one more question related

to this, ⁓ more of a broader question. Do you see any regulatory or ethical challenges that are slowing down AI adoption and maybe even specific to the India market?

Abhisek Nag (35:45)
So, there are, mean. ⁓

With the new data privacy law that’s coming up for us in the country, it might slow down a bit, in my opinion, because you’ll require consent on collecting certain kind of data, which has been, mean, GDPR in Europe has always put that. there will be more today. There’s still how data is handled in India is not, is still not very well structured at this moment. And with the DPDP Act that’s coming up, there’s going to be more…

structured and disciplined use of how data is collected and presented for. So that might have impact on how AI is going to be used because people may have reservations on sharing and then you require consent and all of that. ⁓ And also need to ensure that how the data is used for training purposes for the AI systems is going to be crucial. So. ⁓

The awareness is still low, so it’s not going to be immediate impact, but it’s going to be over the period and back. And I think it’s for the right reasons, because you don’t want AI system which is ⁓ inherently biased. So it’s good that there’s balance of how the data is used and trained. And it creates transparency in the AI systems in the longer run and makes it more open.

Joseph Cole (37:09)
Hmm.

Abhisek Nag (37:12)
and less prone to biases, both algorithmic and ⁓ both data set.

Joseph Cole (37:20)
Awesome. Well, my last question for you is ⁓ let’s look ahead at the future. know, 2020 is only five years away. So it’s like such a big milestone, right? Now, what excites you most about AI’s potential in any people function, HR, TA, talent management, talent development?

Abhisek Nag (37:42)
I see that today, I mean, if you look at… ⁓

about two decades back and I see probably you recollect this and people who are old enough, not the Gen Z’s, but people who are old enough would know this, that the usage of office suite and et cetera used to be a requirement in job descriptions at some point in time. And in future AI is gonna be that hygiene factor. are not, today knowledge of even like two years or even now like knowledge of

particular gen-AI tool or things is a C.V. Brownie point for some roles. But that’s going to be off the table. It’s going to be hygiene factor that everybody is like today. Nobody asks you, do you know how to use an email or Zoom or Microsoft. ⁓

teams or Google Meet, it’s going to be that hygiene factor that everybody will use AI to a certain degree that will become the commonplace norm. And that would, by using that, would elevate the speed of service delivery of HR because a lot of times still today gets on how the people function focuses on internal HR activities as I call it versus employee facing HR activities and by

usage of Gen.ai in the future, is lot of that internal HR task will be much more delegated easily to systems and helping give things to the HR teams to work back with the employees and…

and spend more employee-facing times, resulting in quality conversations with people and elevating the HR function as a whole for organizations.

Joseph Cole (39:36)
Excellent, thanks for that Abhishek. Alright, well, before we close things off today, is there anything else that want to share?

Abhisek Nag (39:47)
The future is we begin that future is bright with JNI. So ⁓ it’s to our advantage whether we start using it now or not. But I’m again slightly open to see how it might shape up. And I was just reading yesterday about Yan Li-Kuan from the head of the AI for Meta ⁓ talking about something on that he didn’t feel that the current stream of JNI

are going to be the forerunner for artificial intelligence as a whole. So if AGI comes in play, then it’s going to be seen how things change. And that’s a different world that we’ll be living in. And I’m quite bullish on that. So let’s see how that shapes up.

Joseph Cole (40:35)
Excellent. How can people get in touch with you? You know, follow you. Okay. Great.

Abhisek Nag (40:39)
LinkedIn, LinkedIn, yeah, connect over LinkedIn.

LinkedIn handle, the global HR.

Joseph Cole (40:46)
Great, well I’m sick, thank you so much.

Abhisek Nag (40:49)
Thank you. Thank you, Joseph. Really appreciate the conversation. And it was really nice, some of the things going back, thinking about AI. I am sure you’d have some use cases. And I would love to know from you also what kind of things that you are doing in AI and how you are seeing the change of AI.

Joseph Cole (41:09)
Yeah, we’ll definitely, maybe we’ll do a follow up on that one. ⁓ But I think that’ll be a good follow up conversation. Awesome, thanks again.

Abhisek Nag (41:17)
Thank you.