Here's How Companies Will Use AI for Recruiting
And why I don't like it.
It feels like yesterday that I got the call from the recruiter after sending in my resume just a few days earlier. They asked the typical protocol questions: What was your greatest achievement at your last job? What do you consider your greatest strengths and weaknesses?
We’ve all faced countless questions like these and can probably recite the answers by heart. After the call, if everything went well, we’d land a face-to-face interview with the potential new boss. Looking back at the entire job search process from a few years ago, it’s clear that human contact was present at every stage. Whether interacting with the recruiter or interviewing with team members, there was always a personal touch.
It was fascinating to see how routine and tedious tasks, like sifting through hundreds of thousands of resumes (which arrived daily for job openings), became streamlined with the introduction of applicant tracking systems (ATS). The first time I heard about this, I was genuinely surprised and had to revise my resume several times to include the essential keywords.
Today, integrating AI into the recruitment process seems almost mandatory. It was bound to happen. But beyond what common sense might suggest, you might ask: Why? The main reason is that companies are eager to incorporate AI across all departments, especially in areas not directly tied to their core business.
Let’s discuss something that impacts all of us: recruitment and job hunting.
My initial take is that human interaction during job interviews is essential and, to some extent, irreplaceable. Despite any inherent biases we may have as humans, the candidate benefits from being able to demonstrate their skills and past experiences through direct communication. Conversely, the recruiter can address the candidate's past challenges in a more empathetic manner and determine if they fit the role. This is the crucial point for me. It’s where the recruiter uses their intuition, developed from reviewing numerous professional profiles.
AI will disrupt this.
AI in the Hiring Process
Let’s talk about Braintrust, the "world’s first" AI recruiter. While this isn’t entirely accurate—since several companies claim the same title (much like Devin trying to be known as the first software engineer)—it’s still noteworthy. Adam Jackson, its founder, explains that this AI is designed to handle the routine tasks of a recruiter 24/7.
This AI can generate a complete profile for a specific position simply by providing an overview of what we are looking to recruit. It can also sort, rank, and select candidates based on variables such as experience, soft skills, and specific skills. Also, the AI schedules interviews to get a more detailed understanding of the candidates.
You can see more details in the following video.
I'm impressed. However, I want to point out a few things that made me question Braintrust's well-executed marketing strategy.
1. Do we want to trust AI with our job opportunities?
One thing I noticed while watching the demo is that this AI can generate a complete candidate profile from just a brief job description. For example: "I'm looking for a full-time senior frontend developer with experience in React, TypeScript, and JavaScript. This position is remote"
The result of this "complete" job description works if you're only looking for technical skills. However, the content is generic and would fit any company. Honestly, we could accomplish this by searching on Google ourselves. What I mean is that this AI misses out on variables like company culture and the specific interpersonal and social skills needed for the position.
This AI starts the selection process by focusing on key variables like knowledge and experience with the specified tools. This means it will search for candidates who meet these criteria, immediately discarding junior profiles. Opportunities for them? None, of course.
Once we've met all the specified requirements, we'll move on to the personal interview, which will also be conducted by the AI. Normally, in a traditional interview, we establish connections through both verbal and non-verbal cues. We might lose that connection for a while. Just like traditional interviews, AI interviews would have their pros and cons. For example, we could fool the AI by simply reading from a prepared script.
All in all, this raises an important question: Do we really want to put our future in the hands of intelligent systems, such as an AI-conducted interview?
2. AI reflects our own biases
We usually think of AI as a technology designed to capture human knowledge while leaving out innate characteristics like bias. However, this isn’t entirely true, especially in the hiring process. Here, what we’re essentially doing is sorting through and filtering out hundreds of thousands of profiles. So, from the start, the AI is learning to establish characteristics or biases, just like we do.
The following chart highlights how AI can exhibit biases based on certain geographic regions. In the study, LLMs were asked to rate their agreement with statements from the Global Attitudes (GAS) surveys and it was found that the models’ responses closely align with those from individuals in Western countries.
Taking into account how these LLMs work, this might lead to AI rejecting candidates in situations where it lacks sufficient information or is influenced by the data it was trained on.
Let’s not be too easily impressed, especially by promises tied to current trends. AI is impressive, but it still can’t match certain human qualities. I like the idea of AI simplifying routine tasks in the hiring process. However, empowering AI for tasks it’s not fully equipped for makes me think that many of us might face frustration in our job prospects.