Navigating AI in Recruitment: Balancing Innovation and Human Insight

“It’s not about the amount we say; it’s about making what we say really count. We can use some of these tools to write the long version, so that we can then quickly create the short version and really dial in.”

– Natalia Bielczyk

What you will learn

• Exploring the impact of Black Swan events on the future of work
• Understanding the role of AI in accelerating job market trends
• Navigating the BANI world with better filtering mechanisms
• Balancing AI and human judgment in recruitment processes
• Emphasizing the importance of work ethic in the AI era
• Discovering personal productivity hacks for focused work
• Fostering empathy and kindness in a technology-driven workplace

Episode Resources

• ChatGPT
• BANI vs VUCA
• Upwork Research Institute
• Artificial Intelligence
• LLMs
• Netflix
• The Coded Bias
• Machine learning

People

• Joy Buolamwini
• Tony Robbins

Books

• Thriving on Overload: The 5 Powers for Success in a World of Exponential Information by Ross Dawson
• Awaken the Giant Within : How to Take Immediate Control of Your Mental, Emotional, Physical and Financial Destiny! by Tony Robbins

Transcript

Ross Dawson: Natalia, it’s a delight to have you on the show.

Natalia Bielczyk: Thank you so much for your invitation. Ross, I’m honored to be here.

Ross: So we have a changing world of work, and people have been talking about the future of work for quite a few years, and I think we’re already well into the future of work, but it’s changing fast. I’d love to start off by just getting your high level perspective on what are the things that we should be looking to in shaping a better future of work?

Natalia: Absolutely. Actually, ever since we faced the Covid 19 pandemic, I think the number of black swan events actually got- I have a feeling that these events got denser and denser, so it’s really hard to I can tell, from a perspective of a neuroscientist that research over the potential future of work is so much more challenging than neuroscientific research because we cannot really foretell in the long run, how these incoming black swan events that we by definition cannot predict, will shape the future of work. Each one of them seems to not necessarily change the future of work, but more like accelerate the progress. So the covid 19 pandemic, it didn’t qualitatively change the the job market, but it speeded up the processes that were already going on by 10 years and and then I believe that the premiere of chatgpt was yet another event that, again, since OpenAI was the first big tech company who showed balls to actually release the top tier software to the public, and then that actually prompted others to come to the scene. That was, again, just speeding up a process that was already going on. Most of these models were already in development for many years prior to the premier of GPT. Seems like one player came to the scene, others followed, and now we have almost an arms race, and that’s fundamentally changing the job market. So we don’t know what comes next. Maybe the US presidential elections will change the scene. Maybe. We cannot really tell, like, what will happen with respect to global events and groundbreaking points in technology worldwide in the next 2, 3, 5, years. We can make some educated guesses for the future. In this episode, I’ll share some of my educated guesses. Obviously, it’s only a guess, but I hope that it’s useful as well. 

Ross: Well, I think it’s also not so much about guessing. I mean, that’s part of the thing: being a futurist, you don’t try to predict, because we don’t know. But it’s around really saying, ‘what is it we can do that can shape a better future’? So there are all these forks in the road and uncertainties, and all sorts of extraordinary things will happen that we can’t predict. I think a lot of it is around saying, ‘well, if we want to create a better future of work, what is it that we need to be doing today’? That’s really the heart of the question.

Natalia: Right. There are a few things that we should be doing as soon as possible. First of all, I think education is always the answer. Let me elaborate on this. At this moment, we live in the world of so -called BANI, which is an abbreviation for brittle, anxious, nonlinear and incomprehensible. Brittle, anxious, nonlinear and incomprehensible world. It’s a new concept, yes, it’s been floating around for the past two, three years, as opposed to the previous concept, the world of VUCA. So now that the world became even more hard to comprehend, that’s the new big thing in the future of work area, BANI as a concept.

Basically, what we need as individuals is better filtering mechanisms. I think it’s really hard to tell the difference between events that we should really be considering as important to us and the events that actually seem relevant. And I have to say that I learn every day, because there are so many global events like the Olympic Games or the elections in the US. I’m Polish, I consider myself European. I often visit the US, but I’m still a European citizen, and for me, it’s really hard to tell sometimes which global events and which news, also in technology, which new software is actually relevant to me. There was a lot of research also, with respect to utilizing AI for work and a new research by Upwork Research Institute showed that 77% of employees, all from executives through specialists to freelancers, they had 5000 subjects in the study, so it was a massive study, and showed that 77% of them actually declared that AI hampered their productivity and increased the workload. 

If you don’t know what is actually relevant and worth using, and you don’t have the right filtering mechanisms, your personal world becomes even more incomprehensible, and your workload and also cognitive load becomes more unbearable than before. I think one thing that we definitely should be doing is to have better filtering mechanisms and become our own Zen masters and the masters of creating, I would call it a sensory warmth, so when we work we have to really be selective with what kind of stimuli we allow to enter to our world, and AI should help us to filter and not give us additional cognitive burden. So that’s something I’m always trying to do for myself. For instance, one way of utilizing AI in our favor while working is to browse using AI as a gatekeeper to keep yourself screened from all the dopamine shots, such as banners online, random advertisements, all the trash that you are actually bombed with every single day. For me personally, browsing World Wide Web through GPT or other LLMs is actually beneficial in that way. The main benefit is not that it’s actually extracted knowledge that is structured to answer my specific question, but it’s more that it screens me from unnecessary stimuli. That I could recommend, using LLMs for that reason.

That’s one thing we can do. But also there’s so many other things. For instance, I think what multitasking is today is often understood as- I think it’s a great misunderstanding, because we always try to do too many things, but our mind only has one channel of consciousness, so we cannot really multitask. We are like computers that are analog, that only have one main process going on, right? They cannot complete multiple processes at a time. So we, like computers, can only switch between different processes quickly. So we are the same, and multitasking in 2024 should be always asking yourself, which task will serve to most of my goals, the highest number of my goals. Choosing the tasks that cater to multiple of your needs, or fulfill multiple of your tasks at the same time, so choosing the most effective tasks and not trying to catch every single task that is on your to do list. Just try to delegate those that are just not as relevant as others. So it’s all about prioritization and filtering. 

Ross: Yes, these are very much echoing the themes of my book, Thriving On Overload, where both on a number of levels, certainly in terms of these ideas of, how do we filter out, or what are the tools we can have to filter out a massive amount of information so that we can make sense of that effectively. And as you suggest, recent neuroscience has demonstrated that people who think they are multitasking are, in fact, cognitively switching between tasks with a significant impact on performance when they do that. So it’s coming back to the organizational structures, that’s something which you’ve studied a lot. We start to think about what happens in organizations if we start to get the productivity benefits of AI used well in many tasks, AI can augment productivity. But it also goes to the human needs of, ‘what do we want from our work and working with others and being in an organization’? So how does some of those do you see playing out in organizations with these potential for higher productivity, but also the fact that we still want to work with other people and not exclusively with digital interfaces.

Natalia: Absolutely, one interesting fact here is that, there is this public perception of AI as being an external threat in organization. The employees- they feel they have to compete on productivity with AI. But in fact, the research also shows that over 80% of today’s knowledge workers are feeling to some extent threatened by AI just flooding the job market. But what often happens is that in organizations, introducing AI to the processes in organizations often the bottom up initiatives coming from employees. And one of the reasons that is not broadly discussed but I think it’s interesting, is that the main reason is because AI does not complain. AI does not have bad days. There’s so many employees out there declaring that they would rather work with GPT, because GPT is not mean. So that’s also why one thing I always tell who I work with, you have to make sure that you’re kinder than AI, because otherwise, it will be your not your bosses, but your colleagues initiative to actually replace you with AI. So that’s something that’s actually quite a common problem, but not often discussed in public, the work ethic. You’re in 2024, every knowledge worker in every organization should be working on their work ethic primarily. The user experience of working with you should be impeccable. That’s one thing to stay relevant, one important factor to stay relevant in the job market. 

Ross: I think the really interesting points there, one is this, Ethan, a researcher in California has suggested that we need to be kind in order to teach the machines. We need to be our best people in order for the LLMs, which are basically just seeing what we do, and that becomes part of the corpus of body, so I think that’s a fair point. And there is also this thing of machines being, as you say, unfailably pleasant, which is why, unfortunately, some people are preferring them to boyfriends and girlfriends, because they never have arguments. They’re always constructive, so there’s some challenges there, but I think these point in a way, to this potential for, in a world of humans, plus well designed AI, that it could help us lead to where we have better behaviors and are kinder and more thoughtful if we start to interact with the machines, and the machines guide us and give us a model, modeling in both ways to create better interactions.

Natalia: Absolutely, I believe that empathy is contagious, so I’m positive, I’m with you. About organizations of today, there was a very interesting documentary by Netflix, authored by Joy Buolamwini, entitled The coded bias. And it was all about biases created by AI in recruitment, AI and machine learning to be, to be precise, in recruitment and in other organization-wide processes. Eventually, any LLM or any machine learning algorithm is just blind to the bias in the data and just to explain briefly how it works, if 99 out of 100 of your software developers are men, then the algorithm learns that being a man as a feature of that candidate is actually supporting the hypothesis that it’s a good candidate because it just fits better the profile of the employees that are already in the organization. There was also research recently on how AI affects recruitment, and it showed that as a recruiter, you can overuse AI. Recruiters that just only use AI to inform them about the ranking for the candidates, and they do not support themselves with any other evidence they actually get worse results as recruiters than those who actually pre filter or just support themselves using AI, but still use their own rationality and interviewing skills to try to spot talent. So being a centaur or cyborg, if you will, is actually a better idea, especially in this, as a gatekeeper to the organization to spot talent, than just using only AI.

Ross: So what does that look like then? So let’s say there is some use for AI in recruitment, because there is a massive number of applications that you need to be able to make sense, there is value in perspectives from that. But at the same time, for many reasons, as you say, including the embedded bias you certainly don’t want to delegate fully to AI. So when you talk about the centaur or the cyborg recruiter, what does that look like? What role does AI play? What role does the human play? How do you bring those together to create a superior recruitment process?

Natalia: That’s a very good question. I think there is no clear algorithm for this just yet. At this moment, what is being done is mostly that AI is used to pre-screen or pre-filter candidates in the first step. So that obviously, especially with today’s ways of applying for positions, with very easy to use tools such as just apply through LinkedIn, sometimes it’s just one click necessary to apply. Easy apply procedures in such circumstances. Obviously, there will be many wannabes who just sit and click, apply, apply, apply. Just meet their daily quota. Okay, today I applied for 100 jobs. I can go to sleep now.

Ross: And everyone with a customized cover letter. 

Natalia: I always tell people ‘don’t do it’. The probability of winning a position this way is just zero. You only create chaos, create trash data, you make it harder for everyone. Don’t do it, be selective. Only apply for positions that are feeling genuinely exciting and you feel like it was written especially for you, for your skills, for your experience, for your personal, professional mission. Otherwise it makes no sense for anyone, for you to apply. But still, that’s always an optimization problem. The easier the application procedure is, the more inclusive it is, because everyone is free to apply, but at the same time you don’t have to have anything, any professional help, any professional tools, you can just click apply, and then you’re in, at least theoretically. But the downside is that the more inclusive the application process, the more noise in the process. I have to tell you that recruiters are one of the most, the most depressed and burnt out professional groups, especially in the times of pandemic, the in the US, the rate of burnout among professional recruiters was 96% it’s extremely hard for them to deal with this cognitive burden so many applications coming in. That’s why, in the pre-filtering stage, using algorithms that just simply compare text between applications and job offer, they are meant to just filter out these random applications. Sometimes, unfortunately, it’s just throwing out a baby together with the water, unfortunately. And sometimes you’re losing also good applicants. But for a recruiter at this stage, it’s like they’re… 

Ross:  It’s impossible to go through everything. So you do need some system at some point.

Natalia: Yes. 

Ross: So, where are the points where humans augmented by AI can help make better decisions in a recruitment process? I mean, where it is a human first decision, but AI assists it.

Natalia: At this moment, one area where AI is useful in recruitment is to help hiring managers design better recruitment questions that are more personalized per candidate. So hiring managers can better prepare for interviewing the top candidates, and they can easier analyze their professional story, adjust the questions to the candidates professional past, and this way they can learn more. So now, with very little amount of additional work, they can actually get better conversational interviews and get deeper into the candidates motivation and personal story, other than just asking a few standardized questions from the list, and just having the checkpoints down. I think the quality of interviews actually got higher due to AI and vice versa: the candidates can better prepare now for the interviews, they can let the LLM screen the employer’s website and learn more about their mission, and learn more specifically about the recruiter who is going to interview them. Now, if you really wish to prepare and you’re willing to spend time, you have all the means to do so. It became more fair, in a way, because even without having resources to hire a professional coach, or even without much experience in the job market, using AI, you can much better prepare and get a much better chance of getting in.

Ross: Yeah, one of the very nice use cases for LLMs is to practice job interviews, where you train the LLM to be very close to what you expect the interviewer to be and to be able to practice live in a fairly realistic situation.

Natalia: However, I would still advise to still have a chat, even with a friend. Practicing with GPT or other LLMs is a good idea, for sure, but also still I often see that people are perfectly prepared in terms of content, but still, when it comes to human interactions, they don’t look in the eye. They just look around, they look to the ceiling. They just didn’t have that practice in front of another human, even on screen. So I would still encourage whoever is listening and just in the process of applying for jobs to still practice with another human, because it’s the same case as the famous presidential debate between Nixon and JFK, right? The way you present yourself on screen, even if your content is better, but you look sweaty, you look like you don’t know what’s going on, you will eventually lose the competition to those who look relaxed, right? It’s really important to have that human interaction too, and practice in front of another human

Ross: Yeah. I’d love to dig into some of the ways in which you maximize your productivity, your capabilities, your potential, using AI or other tools. I mean, what are the things which you have found effective and useful, 

Natalia: Right. One thing I have to say, first of all, is that the whole notion of productivity is not very well defined. I myself, I have to say that I’m guilty of spending way too much time at school. I did complete three master degrees, and then I went for graduate school, and I literally woke up when I was already in my 30s as a person who never, ever had any real life experience outside of my school desk and when you’re at school in an education system, you have a top down way of assessing your productivity, right? You go for tests. You go for exams. They tell you, you did well or you did not do well. I had to wake up in my 30s and now I was on my own, and I started a company.

What’s your productivity if you just spend your day working, but you didn’t sell anything today? You didn’t close any deal either. You just keep on going with your daily tasks. How do you assess if you’re productive or not? So it was really hard for me to, like, reconceptualize what productivity means and I eventually came up with the idea that, okay, I have my own to-do list. Nobody can do it for me. I have to set my own goals. And there are only two categories of days. There are days which I win and days which I lose if I stay on the track and I don’t let the rest of the world hassle me out of my goals, for instance, by just dragging me to the TV screen because something happened. There are lots of big events these days that can easily just pull you out of your bubble, and just drag you to the TV screen. So if I don’t let the rest of the world hassle me out of my way, and I still stay on track, that’s a win, and if I get distracted, that’s a loss. That’s the only, that’s the only criterion that I use right now.

About my personal ways, well, I think it’s a combination of state of the art methods for productivity, even games from the whole standard literature, even Tony Robbins and even these old tricks that are really good for productivity still. Actually Tony Robbins’ “Awaken the Giant Within” was the first book I ever read, because when I was three years old, my mom was obsessed with Tony Robbins. Just read it to me without my consent. But that was effectively the first book I ever read. But I also think that if you want to have results better than the most, and as a matter of fact, most people never get to realize their professional dreams. They always have these what-ifs on their mind. So if you want to achieve better results, you have to use other methods, you have to try something. That’s my belief. And I always enjoy finding my own tricks that are original, and I have never, ever read about them anywhere. For instance, when I was an undergrad student, I used to study three majors all the time, and I always had lots of exams in the exam period. But I learned that if I condition my brain, if I wear black T shirts to every exam, and I never wear black otherwise, I condition my brain to always feel like I’m on the fight mode once I dress black, and I told my friends, if you see me in black, that means I’m going to the exam. Don’t talk to me, don’t distract me, because I’m going to fight. And it really worked for me. So every single time I was wearing black, I was like, really, let’s go. This is my day to shine. And I was laser focused on the task. So this was something I just found out for myself. Or, for instance, when I record educational materials, like courses, like it’s always better, I always enjoy more when I talk to somebody than if I just talk to the screen. So I put my favorite teddy bear behind my laptop and I talked to that little guy, right? I’m like, Listen, this is how you do it. And then I really feel like I have so much more energy just talking to the teddy bear than just talking to the wall. So these are the things I especially enjoy, those little tricks, little life hacks, because first of all, it’s a sign of creativity, and it works. Little tricks that make a lot of difference when you just integrate, like, there is a compound effect over lifetime of these little everyday tricks. And we are all like hackers, right? We all have our own ways to solve problems, but I always encourage people also to just find your own productivity hacks that work for you that nobody ever found. It’s just your own way to solve your own problems. And yeah.

Ross:  Absolutely. Your idea of the teddy bear is similar to what software programmers use with the rubber duck, tape a rubber duck on their screen. If it gets wrong, they have to explain it to the rubber duck to clarify their own thinking.

Natalia: Right. But my duck has to be fluffy, so I prefer teddy bears.

Ross: I think that’s very reasonable. Where can people go to find out more about your work and what you do?

Natalia: Well, so I think the best way is to follow on social, because I put a lot of content. Some of this is just reviewing the recent research on productivity, future of work, the current state of the art in a job market. So follow me on LinkedIn, follow me on Twitter or X, it’s a good idea. I also have a little YouTube channel, but it’s kind of on a stealth mode right now. I just think that social media just works best because I also do a lot of initiatives, like organized conferences and so on, it’s just everything I communicate to public, so I think this is the best way. Yeah, thank you. 

Ross: We’ll share links in the show notes for the podcast. So thank you so much for your time and your insights today. Natalia.

Natalia: Thank you so much. Ross, it was a pleasure of pleasure on my side.