Hired or hidden? AI’s new power in the job market

Summary

Business Daily — BBC World Service
Episode: Hired or Hidden? AI’s New Power in the Job Market
Aired: May 5, 2026
Host: Ed Butler
Guests: Daniel Chait (CEO & Co-founder, Greenhouse), Professor Elizabeth Keelan (King’s College London)

---

Episode Overview

This episode of Business Daily investigates how artificial intelligence (AI) is reshaping the job recruitment landscape. Ed Butler and his guests delve into the increasing role of AI in both applicant screening and job seeking, highlighting the unintended consequences this technology brings for job hunters and employers alike. The discussion examines whether AI’s influence is democratizing opportunity or exacerbating tensions, bias, and dissatisfaction on both sides of the job market.

---

Key Discussion Points & Insights

1. The Bleak Reality for Job Seekers ([01:43]–[05:13])

• Statistics reveal a tough market: U.S. graduate unemployment soared to 5.6%, and UK graduate hiring by top accounting firms is down by 44%, with AI cited as a key driver.
• Real-world experiences:
  • Karina, a highly qualified Ukrainian master’s student in the UK, describes applying for over 400 jobs, facing multiple AI interviews, and feeling filtered out before human consideration ([03:52]).
  • Another UK graduate voices discouragement at being repeatedly “scouted and screened out” by bots ([05:13]).
• Quote:
  “I had five interviews with AI...sometimes I don’t even hear the answer back and I feel like AI filters me out and a real person is not even being able to look at my CV.”
  — Karina ([03:52])

2. The "AI Doom Loop": Escalating Arms Race ([05:57]–[08:41])

• Daniel Chait explains both applicants and recruiters now use AI: applicants use tools like ChatGPT for mass applications, while employers use AI to wade through the flood of CVs.
• This cycle—AI-enabled application flooding leads to stricter AI filtering—worsens the experience for all.
• Quote:
  “We’ve referred to it at Greenhouse as the AI doom loop...each side is kind of using AI to solve what they see as their own problem. But...the more AI that’s been used, the worse the problems have become for both.”
  — Daniel Chait ([07:23])

3. Is "Gaming the System" Cheating—or Adaptation? ([08:12]–[10:50])

• Professor Keelan argues candidates are acting rationally, using AI to maximize opportunities. Employers, meanwhile, see outcomes as “cheating” due to generic, AI-generated applications.

• There is growing ambiguity regarding what is considered “okay” use of AI on both sides.

• Quote:
  “Candidates are playing the system, they’re gaming it like you would do as a rational actor in the market...I don’t think that serves anybody.”
  — Professor Elizabeth Keelan ([09:09])

• Chait points out the double standard: companies use AI throughout recruitment, yet often judge applicants for doing the same. The “rules” are unclear and rapidly evolving.

4. AI & Bias: The "Black Box" Problem ([10:50]–[13:23])

• Professor Keelan highlights AI’s potential to amplify bias, e.g., penalizing candidates for accents or non-standard English ([10:50]).

• Even when systems don’t intend prejudice, their training data or parameters can exclude worthy candidates.

• Lack of transparency means job hunters often never know why they’re rejected.

• Quote:
  “It could be, for instance, that these people have an accent...the AI system has been trained...not to recognize these accents. So for them the error quote is going to be higher and they’re less likely to progress forward.”
  — Professor Elizabeth Keelan ([10:59])

• Chait argues that many recruitment problems—bias, lack of feedback—pre-date AI. The challenge is whether AI will perpetuate these issues or be used to improve transparency and opportunity.

5. Opportunities: Rethinking the Process ([13:23]–[15:39])

• Chait envisions potential for more inclusive hiring: AI-driven screening can allow more candidates “in the door” (e.g., automated initial interviews), potentially reducing certain forms of bias.
• Current dissatisfaction is shared by both applicants and employers—the process is seen as impersonal and unreliable.
• There’s a need to redesign hiring to be more human and trustworthy, rather than accelerating the same ineffective processes.
• Quote:
  “Are we going to magnify the same problems that human recruitment has been having forever, or are we going to consciously design AI to overcome those problems and think differently?”
  — Daniel Chait ([13:23])

6. Towards "AI-Proof" and More Human Hiring ([15:39]–[18:53])

• Employers are experimenting with ways to assess candidates more authentically:
  • Greenhouse’s solution: let candidates use AI to prepare, but conduct live collaborative interviews to judge genuine skills ([16:55]).
• Danger of regression: Butler contemplates whether overwhelmed companies will revert to “nepotism” (hiring through personal contacts), but Keelan strongly warns against this, emphasizing merit and diversity.
• Collaboration and adaptability are highlighted as skills for the future—AI isn’t an excuse to return to outdated practices.

7. Advice for Job Seekers: Quality over Quantity ([19:01]–[19:44])

• Chait warns against “spray and pray” AI application tactics, advising instead:

  • Focus on networking, personalizing applications, researching potential employers, and crafting compelling notes.
  • Despite market challenges, these traditional approaches remain most effective.

• Keelan adds: encourage graduates to stay selective, seek roles where they can excel, and avoid pursuing every opportunity aimlessly.

• Quotes:
  “Generally speaking, our data shows [‘spray and pray’ with AI] doesn’t work. What does work is good old-fashioned networking, research, personalization...”
  — Daniel Chait ([19:01])

---

Notable Quotes & Memorable Moments

• “It’s not only that AI is going to take your job, it’s also going to not select you for a job.”
  — Professor Elizabeth Keelan ([04:23])
• “If everyone can have a conversation with a company because they’re able to automate those initial screening calls, you can eliminate entire categories of bias.”
  — Daniel Chait ([14:29])
• “Returning to a world of nepotism...I would hope that this is not a solution to the crisis that we are seeing.”
  — Professor Elizabeth Keelan ([18:16])

---

Key Timestamps

• 01:43 — Introduction to the AI hiring landscape & job market stats
• 03:52 — Karina’s job hunt struggles with AI interviews
• 05:13 — UK graduate shares discouragement with AI “bots” in recruitment
• 07:23 — Chait introduces the “AI doom loop”
• 10:59 — Keelan explains AI and bias (language/accent example)
• 13:23 — Optimistic view: can AI fix what human bias couldn’t?
• 16:55 — Greenhouse’s approach to “AI-proof” hiring
• 19:01 — Practical advice for job seekers
• 19:44 — Closing thoughts and encouragement

---

Tone & Language

The discussion remains thoughtful and candid, occasionally blunt, as both guests grapple with the real struggles AI has introduced for both companies and job seekers. There’s an undercurrent of concern about depersonalization and bias, balanced by hope for future improvements if AI is designed more conscientiously.

---

Bottom Line

AI’s dominance in recruitment is creating widespread frustration and unintended “arms race” dynamics. While technology can process vast applicant pools and fight some forms of bias, it often heightens the sense of impersonality and hidden barriers for job seekers. The current cycle—more AI on both sides leading to more filtering and less satisfaction—is unsustainable but potentially reversible. The clear advice: both job seekers and employers need to rethink their approaches, returning to genuine human connection and careful design if they want AI to serve, not stymie, the job market.

Transcript

[A]

This BBC podcast is supported by ads outside the uk.

[B]

Today's episode is sponsored by Smart Travel, a podcast from NerdWallet. Do you have a friend who treats airfare like a puzzle and somehow always ends up with a first class upgrade? Smart Travel is like that friend, minus the group texts. NerdWallet's travel journalists break down the pros and cons of travel decisions like when lounge access is worth paying for and and which hotel loyalty programs actually pay off. Make your travel dollars work harder. Follow Smart Travel on your favorite podcast app. Ever invest in something that seemed incredible at first but didn't live up to the hype? Like those five dollar roses at a gas station? Or a second hand piece of technology that breaks in the first ten minutes. Marketers know that feeling. We optimize for the numbers that look great, impressions reach and reacts. But when they don't show revenue, well, that's a not so great conversation with the CFO. LinkedIn has a word for that. Bullspend. Now you can invest in what looks good to your CFO. LinkedIn Ads generates the highest roas of all major ad networks. You'll reach the right buyers because you can target by company, industry, job title and more. So cut the bull, spend. Advertise on LinkedIn, the network that works for you. Spend $250 on your first campaign on LinkedIn ads and get a 250 credit for the next one. Just go to LinkedIn.com Broadcast. That's LinkedIn.com Broadcast. Terms and conditions apply.

[A]

Hi there, I'm Ed Butler. Welcome to Business Daily on the BBC World Service, where today we're going to be engaging in what you might call a double take, exploring different perspectives surrounding one major issue. Today our theme is AI. Sure, we know that artificial intelligence is irreversibly changing the world of work. Jobs are going to disappear, other jobs will be created. But today I'm looking to understand how AI is disrupting recruitment, specifically because from what you'll hear soon, the experience for job seekers has become pretty bleak. Is AI helping the job search? That's Business Daily from the BBC. Yes, today I'm going to be looking at all of these subjects around AI and specifically recruitment. And I'll do it in the company of Daniel Chait. He's the CEO and co founder of Greenhouse, a recruitment AI software firm that helps companies in the US hire the right people. Hello there Daniel. How are you doing today?

[C]

Hi Ed, Great to be here.

[A]

And Professor Elizabeth Keight, Helen from King's College London, she's with us Too. She studied developer bias in AI software tools. Hello, Elizabeth.

[D]

Hello. How are you doing?

[A]

Very good indeed. Thank you both for being with us. Let's get into this some numbers. Did you know the unemployment rate for U.S. college graduates has soared to 5.6% at the end of last year? This, according to the Federal Reserve bank of New York. That's outstripping the general unemployment rate by some margin. There are other red flags as well. Within the past year in the uk, the four accountancy firms who usually hire a lot of graduates, they've cut their recruitment of graduates by as much as 44% in both cases. AI seems to be the key in terms of the reduction of hires. It's a reality facing graduates around the world. Of course. Karina is Ukrainian, but she's completing her master's in governance in Cardiff in the uk. She's already had work experience at the United nations and. But she's been rejected, she says, for hundreds of job roles in the uk.

[E]

I've been looking for a job since the end of February and by this point I applied for more than 400 jobs and I had five interviews with AI. And this is the most confusing part, because sometimes I don't even hear the answer back and I feel like AI filters me out and a real person is not even being able to look at my cv.

[A]

Professor Keelan, does that ring a bell with you? This is a woman who. She's speaking to nothing but bots when she's going through the application process.

[D]

Yeah, I think that's a very common experience. Certainly the students that I teach report that many of the interviews that they have are effectively with a machine. Right. So when they're asked automated questions, however, I always remind my students that it's not automatically that the AI assesses them without human input. Very often these interviews are assessed by human people. Right. Assessors who actually are trained to help candidates shine. But I totally get the sentiment that it's not only that AI is going to take your job, it's also going to not select you for a job.

[A]

Yeah, indeed. And it feels depersonalised, doesn't it? I guess that must be demotivating for a lot of applicants. Carina is not an isolated case. The BBC has also heard from other job hunters struggling to get past AI gatekeepers.

[F]

Companies like to use AI agents as a form of speeding through job applications and looking at candidates and reviewing whether they're likely to fit the company or not. Although from a corporate sense it might feel amazing. But for US graduates and other side of the process, it's not really encouraging, especially when you go through that process, typing out your motivations for the job and things like that, just to see at the end, like your application will be under review by the likes of bots.

[A]

That is another UK job seeker. Bots, he says, basically scouting and screening them out every time he goes through the application process. Daniel Chait, listening to that. You work with AI tools to help companies. Are UK graduate experiences unique? Are you seeing that in the States as well?

[C]

Yeah, look, I think these trends are common across the globe in many different hiring markets. Even going back to 2020, when the sort of move to online and remote work happened, all of a sudden everybody went from being able to apply to just a few jobs in their local area that they could commute to, to being eligible for potentially millions of jobs anywhere. And so application volume started to go up. And then with the end of the zero interest rate environment in late 2022, application volume skyrocketed even more. At Greenhouse, our customers collectively collect about 22 million job applications every month. So we see lots of hiring data. And what we've seen is that over the last several years, applications per recruiter are up over 400%. And that's even before AI gets into the picture. There's a lot of trends in the hiring market that almost predate AI that have made it really hard to find a job.

[A]

Is it before AI? I mean, I'm interested because I was seeing similar numbers. 11,000 job applications made every minute on LinkedIn. It's not just recruiters, of course, using lots of AI tools to screen applicants, but it's the applicants themselves using ChatGPT or whatever AI tool they have at their disposal to grapeshot the entire market with applications all over the place. Using AI to kind of manufacture an application.

[C]

Yeah, we've referred to it at Greenhouse as the AI doom loop, which is that, you know, everybody's using AI to solve their own challenges. Candidates are trying to get jobs, and so as the market moves against them, they use AI, as you say, using ChatGPT and other type of tools to sort of automatically apply to more and more jobs. Kind of spray and pray. Well, that fills up everybody's hiring pipelines with more and more applications. And so companies then in turn turn are using AI to try to filter it down automatically and get through all that noise and find the real applicants they need to hire, which in turn makes it less likely to be seen by a person. And so candidates send out more and more auto applications, and so. And so the Cycle goes. And so each side is kind of using AI to solve what they see as their own problem. But in doing so, the more AI that's been used, the worse the problems have become for both. It's really a bad situation, a doom loop.

[A]

Elizabeth it wasn't supposed to be like this AI was supposed to be making everything better and more efficient, wasn't it?

[D]

Well, and it clearly is helping with that. Right. So as a candidate, you try to maximize your opportunities. So, you know, now we have the opportunity to also use AI to make your answers better and make it resonate with what the employer wants. So obviously candidates are going to use that, and that is met by employers who want to hire the best fitting candidates. And obviously with that comes, you know, the challenge of how to select them, how to find them.

[A]

Yeah, but I mean, in a sense, isn't that just from an employer's point of view? I'm going to put this kind of more brutally. It's kind of lying. I mean, people are just using technology to manufacture what looks like a very gorgeous, glossy and perfect application. It's not real, it's not really them. And obviously the employer wants that because that's the person they want to hire. So consequently, the system has actually made things worse. It's made things worse both from the employer's perspective and from the applicant's perspective.

[D]

Yeah, the candidates are playing the system, they're gaming it like you would do as a rational actor in the market. And I know that employers very often see that as cheating because they want to hire people with certain skills and characteristics. And what they get is instead a generic AI generated response. Right. And I don't think that serves anybody. And I think it's purely a function of the volume that we see at the moment.

[C]

And look, I want to be, I want to be really clear. The job seekers aren't the bad guys here. The stories that we heard earlier and the this piece are very real. And people are trying to get jobs. And even as it's gotten easier and easier to apply to jobs, it's harder than ever to get a job. And so they're trying what they can and they're using the tools that they have. And I don't necessarily blame people for sending out lots of job applications. I think when you get into these questions that you're asking about, like if people are using AI to tailor their resume to the job or to help them answer interview questions, now you get into these really sticky situations of like, what's actually okay. And I think from Job seekers perspective. What they're saying is, look, if you hire me into today's workforce, you're expecting me to use AI. You want people that are going to be fluent in these tools. And by the way, you as a company are using AI to post job descriptions and to help interview candidates and automate lots of your processes. Why is it okay for you to use AI, but for me as a job seeker not to? And I think the truth is the rules about what you are and aren't allowed to do at various parts of the job seeking process are very unclear and they're changing constantly. So it's easy to point the finger and say, oh, they're cheating on a job interview. And surely that happens. But I think the real issue is how is it that companies are able to identify the skills that they need for tomorrow's workforce, which is absolutely AI enabled, without having job seekers use AI at all? It's not clear.

[A]

Professor Keenan, just quick word. You looking at biases within AI, how is that working? What have you seen?

[D]

I think that's partly what we also heard before in the responses from the candidates. They're being filtered out and they don't really know why. So it could be, for instance, that these people have an accent and model. The AI system has been trained in a way not to recognize these accents. So for them the error quote is going to be higher and they're less

[A]

likely to progress forward because AI is hearing them speak bad English.

[D]

It thinks, no, it's not the bad English, it's just that trained on a specific form of English which might be American English or it might be a certain form of British English that these machines are being trained on. And once you have an accent, the machine is less able to understand you and then you have more errors in your data. And that's not necessarily a problem in itself if you have trained the system very well. Right. So you can deal with that and you can ensure that candidates will get through.

[A]

You're listening to Business daily on the BBC World Service.

[G]

Energy Trust of Oregon, we know it isn't easy. The tremendous costs of maintaining a home, living in a house with drafty windows, poorly insulated spaces, or equipment that's running well past its prime. But we also know how to help you upgrade your systems, insulate your spaces and replace what is old with new energy efficient solutions. Find cash incentives at Energy Trust. Energy Trust of Oregon. More power to you. When you manage procurement for multiple facilities, every order matters, but when it's for a hospital system, they matter. Even More Grainger gets it and knows there's no time for managing multiple suppliers and no room for shipping delays. That's why Grainger offers millions of products in fast, dependable delivery so you can keep your facility stocked, safe, and running smoothly. Call 1-800-GRAINGER click granger.com or just stop by Granger for the ones who get it done.

[A]

My name's Ed Butler. In today's program, we are looking at job recruitment and whether AI is helping or hurting. Let me turn to Daniel again. Now, we've heard some of the problems so far. I mean, how do you see the solution to that? I mean, what Elizabeth, for instance, just spoke about the idea of a kind of black box within AI that could be basically driving prejudice or bias in some way.

[C]

Yeah, you know, I don't disagree with anything that Elizabeth pointed out, but I think there's a way to look at those similar facts and point to a very optimistic vision of the future for AI. First of all, decision making and hiring has always been extremely biased. There's been problems with inconsistency, with bias in hiring as long as there's been hiring. And secondly, the complaint that job seekers are lacking transparency. They never hear back. They get ghosted. These have been happening for decades. You can look at any candidate experience survey for the last three decades, and they all show the number one complaint that job seekers have is they never hear back and they never get any reasons. And so these things, again, they all predate AI. And I think the challenge and the opportunity that I presents is are we going to magnify the same problems that human recruitment has been having forever, or are we going to consciously design AI to overcome those problems and think differently? And I think that's where the opportunity lies to say, well, hang on a second. If you take the interview that we played at the beginning of this, of this segment, the candidate that's being interviewed by an AI instead of by a person, chances are that they wouldn't have gotten an interview at all. Because most jobs today have hundreds or thousands of applicants for one open spot, which means you've got a very, very small chance of having a conversation at all. But if everyone can have a conversation with a company because they're able to automate those initial screening calls, you can eliminate entire categories of bias because the very idea of who qualifies at all for an interview simply goes away.

[A]

Is this an arms race between applicants and recruiters? I mean, is that how you see it?

[C]

If you talk to job seekers today, they see the hiring process as A system to beat. They see it as not working for them. And if you talk to companies, you know, they're drowning in resumes and they don't trust, is this really the person that they say they are? Is this really the work that they've claimed on their resume that they've actually done? Like, I agree with you that it's almost a unique time where neither side is happy with how hiring is going. But the way forward is to think about, you know, not just how do we continue to turn the same cycle faster and faster and use AI to, you know, make it. Make volumes go up and the experience go down, but rather this is an opportunity for us to pause and reimagine the process altogether in ways that I think can be a lot more human and can bring a lot more trust.

[A]

Professor Keelan, do you imagine a more human system? Is there an opportunity here?

[D]

I think what people perceive as a lack of human system is effectively, you know, making the human and the design process invisible. Right. So that kind of echoes the question of, you know, should we bring humans back in? Would it change if suddenly humans would start to interview? And yes, as Daniel has said, our, you know, human recruiters, interviewers are biased too. But similarly, the systems themselves are biased in many different ways that are very hard to spot. And very often we also need to consider that these systems are not neutral. They are not built outside of human interference. They are already built by humans. That people design those systems. People are involved in evaluating and making those judgments stick.

[A]

Daniel, I have been reading several companies talking about this problem. On the one hand, as you say, they're drowning in resumes, firing at them. They're probably feeling a bit demoralized. And I'm wondering what the companies themselves are doing in response. I mean, obviously more human interviewing processes, perhaps more surprise. I don't know if there's such a thing as an AI proof screening process anymore, because AI is very sophisticated.

[C]

Leading companies are thinking about how do they do AI proof hiring. And so I'll give you an example that we've done at Greenhouse, where historically we relied on lots of case studies. Which was the best science in hiring told us was the best way to identify whether the person had the skills to do the job. But increasingly, if you send somebody some work to do at home, whether it's a case study or some other problem, they're going to use AI and you're not really going to understand who they are. So we've shifted that process to be much more of a you can use AI Ahead of time, we'll send you the case study and you should use AI to help you prepare and research and then you get on live. We're going to do some collaborative work together as part of that interview. And so we really get a sense of who you are and how you think. And so companies are very much reexamining the hiring process to figure out how can you crack open who that person is and understand them best and let them show you who they are.

[A]

I mean, if you are a small to medium sized firm saying to themselves, look, it's not working, I am getting bombarded with basically fake resumes which I don't trust. I can't sift through all of this noise to get to the real applicants here. Maybe I should just return to, I don't know, good old fashioned nepotism and ask my best friend if they know anyone who'd like a job. Because actually, I might actually save myself an awful lot of time and get to the best person. In other words, this arms race might force people to completely drop the system and just go back to something very, very simple.

[D]

I would hope that this is not a solution to the crisis that we are seeing. Returning to a world of nepotism and getting a job because of who you know rather than what you know. Because we have seen over many decades that, you know, the quality of results improve vastly if we have more talented individuals who are in those posts. And we know that, you know, nepotism is rarely producing those systems. So instead let's think about how can we encourage candidates to shine, right? And we see that in our classrooms where, you know, students are AI first and certain things, you know, AI can do fantastically for them, but we increasingly get them to work together to collaborate.

[A]

Daniel, quick word of advice. You're an applicant going for a job now, maybe a graduate applicant. What's your general tip to them?

[C]

Yeah, I think there's a lot of temptation to sort of spray and pray. You buy one of these tools and it promises you to use AI to automatically apply to thousands of jobs on your behalf. Generally speaking, our data shows that doesn't work. And what does work is good old fashioned networking, research, personalization. Spend time getting to know the companies you want to work at, find out what's important to them, write compelling notes and emails, and then, unfortunately, the state of the market means it's a lot of hard work. You have to do a lot of those in order to get ahead. But that's by far the best approach.

[D]

Professor Keelan, if you think about it very carefully. Graduates have very limited experience. That's the only experience they have of the job market. Right. So we need to ensure that they keep engaged and really pick the jobs where they can shine rather than going for everything.

[A]

Okay, we'll leave it there. Thank you both very much indeed. We've been looking at AI and the recruitment process. I hope you enjoyed it. My thanks to Professor Elizabeth Keelan from King's College London, to Daniel Chegg, the CEO and co founder of Greenhouse, a recruitment AI software firm in the us to both of you and to my listeners out there, thanks all for listening. We'll be back soon with another edition of Business Daily. Take care.