Hiring in the age of GenAI

As a hiring manager, I’ve recently started the process of recruiting for an open role on my team. This post is not an advertisement for that role but rather an opportunity to share some loud thoughts and reflections after my initial experiences with this hiring cycle.

To set the stage: the last time I actively hired was two years ago—before the widespread adoption of GenAI tools across industries. Over the past couple of weeks, while screening candidates, So far, 75% of the people I interviewed seemed to use GenAI tools to answer questions during the interviews. Furthermore, I suspect that many of these candidates also presented fake profiles.

Let me be clear: this is just a theory, and I can’t prove it definitively. However, I found the experience shocking, not only because of what was happening but also because I struggled to understand the intentions behind it. Writing is thinking for me, so I decided to put my thoughts on paper.

It’s worth noting that the candidate sample size has been small so far. Still, I’m so surprised by what I’ve observed that I felt compelled to share this story.

Disclaimer: I used ChatGPT to edit this post :) Historically my wife would be kind enough to edit my posts, but ChatGPT saves a lot of time for both us nowadays.

General characteristics I’ve observed

Here are some patterns I’ve noticed among candidates:

  1. Buzzword-laden introductions: Candidates start with a nonstop stream of buzzwords, saying things like, “I did X, which improved Y by Z percent.” While this existed before AI, there’s now a complete lack of authenticity—it often feels like someone reading from a script.
  2. Superficial knowledge of listed companies: Candidates claim to have worked at well-known companies but can only describe generic roles and teams. They struggle to go deeper into functional areas, domains, or the organization’s landscape. Sometimes, they can’t even explain the company’s business.
  3. Inability to answer questions about their CV: Candidates often repeat vague statements about their work without elaborating. When I probe deeper, they simply repeat the same points without offering any meaningful detail.
  4. Easily manipulated responses: By suggesting a flawed idea (e.g., “Maybe X approach is better?”), candidates will agree enthusiastically, even if the suggestion is objectively terrible.
  5. Claiming expertise in everything: Candidates confidently answer every question and claim to have experience with any technology or scenario I mention.
  6. Lack of meaningful questions: Candidates struggle to ask thoughtful or relevant questions themselves, often posing very simplistic or even nonsensical queries that don’t align with the role or the discussion.

The puzzling intent

What puzzles me most is the intent behind these behaviors. I can’t imagine how such candidates expect to pass a reasonable hiring pipeline. Is the goal to gather intelligence on interview processes? Is it desperation, leading them to cheat because they feel they have no other option? Honestly, I have no idea.

Identifying patterns and signals

I’ve started to identify patterns and signals that can help me screen out such candidates early in the process.

During CV screening

  1. Missing LinkedIn profiles: Candidates list companies and roles but don’t include an active LinkedIn profile or any other online presence.
  2. Untraceable email addresses: A quick search of their email address yields no results online.
  3. Big enterprise experience without specifics: Candidates list large, anonymous organizations where it’s hard to verify their roles or contributions.
  4. “Spaghetti” experience: CVs list a broad, mismatched collection of programming languages and frameworks that seem to cover everything but lack depth.
  5. Too-good-to-be-true matches: CVs magically align with the exact criteria and requirements of the job posting.

During the interview

  1. Autorobotic answers: Candidates don’t listen to questions properly and give generic or rehearsed answers.
  2. Repetition: They repeat the same answer for different questions, often quoting sentences directly from their CV.
  3. Long pauses: Candidates take long pauses before answering, not for thoughtful reflection, but seemingly to buy time while repeating random filler phrases.
  4. Superficial understanding of their CV: They cannot answer specific questions about their current or past roles, teams, or company.
  5. Nonsensical questions: They ask simplistic, irrelevant, or nonsensical questions that do not reflect any effort to engage with the role or the discussion.

Be bold and ask directly

If I suspect that a candidate is using AI during the interview, I am going to ask them directly:

  • “Are you currently using AI tools to help you answer these questions?”
  • “Can you tell me more about how you prepared for this interview and whether you used any AI to assist you?”

I think being upfront can often reveal interesting responses and gives the candidate an opportunity to clarify their approach.

If their answers seem evasive or insincere, I am going to poke further by challenging their claims with deeper questions or even introducing deliberately flawed ideas to see how they react.

And if the suspicion persists, I am just going to stop the interview early. It’s better to invest time in candidates who genuinely align with the role and bring authenticity to the process.

Final thoughts

Hiring has always been challenging, but the rise of GenAI adds a new layer of complexity. As hiring managers, we must adapt our processes to detect and handle these behaviors effectively. At the same time, we need to keep an open mind and ensure we’re not overgeneralizing or dismissing genuine candidates.

Written on December 19, 2024