Everyone's using AI to get hired. Now what? | BBC News transcript

Looking for a job is hard, but is AI making it harder? >> Feels like I'm talking to a robot. It feels really disheartening. >> The survival of the fittest now. Really? >> AI stands accused of [music] ruining the job market. >> Did you buy that? >> No, I don't. >> But has a doom loop been created, where AI is used to write more and more [music] applications, and where employers increasingly use AI to sift through them? >> Two out three of our clients are really struggling with the volume. >> They can simply go click, click, click.

If you do get through the AI short listing, guess [music] who might be doing the interview. >> For once, I had this AI tell me that I lacked empathy. >> It's like you're speaking to a robot essentially. And [music] I don't think that's >> Well, you are. >> Well, you are exactly >> recruitment is being rewired by technology. >> But is it better either for the applicant or for the employer? To help understand the issues, I wanted to hear from people in the middle of this.

Kate runs an AI consultancy. Matt is a recruitment executive and Neve, Addie, and Santina are job hunting and sending a lot of applications. >> Over 200 now, about 50. >> It's been 100 so far. That's just in the last year. >> This is part of an application explosion which is happening for several reasons. Competition is fierce. The number of vacancies is down. The number of graduates is at a record high. And AI is helping to write more and more applications. 20 years [music] ago for graduate vacancies, there were on average 38 applications per job.

Last year it was 140. Let's look at this from both sides. The employers [music] and first the job seekers. You're trying to get a job at the moment. You're finishing university. How central is AI to to doing that? >> It's really central. Um I I use it now for finding the jobs itself. When I'm searching for jobs, I have my automated agent that goes scrapes through the whole internet seeing which roles apply are opening, which ones the deadlines are, and it helps me narrow the the selection down. >> Maybe put in keywords, see which ones can make sound better.

Even when I know I want to portray a certain side of myself and what I want to say and AI can help me sound better, maybe >> edit my CV, I would make sure that it has the the same words that there are in the job description. >> Addie, Neve, and Santina aren't alone. One survey found 73% of students and graduates use AI at some stage during the application process. I'll be honest, I'm surprised it's not even higher. After all, AI is very helpful.

And there's one part of the process where it's being used a lot. Keywords. >> Job ads ask for certain [music] experience and attributes. AI can make sure your application includes those words. That's particularly helpful when an AI sift is looking for them. But here's something that might surprise you. Using AI to shortlist people isn't as widely used as you might expect. One survey found that 75% of candidates believe AI was being used to screen CVs.

The number of recruiters who reported using it was 21%. On AI sifts, the UK's Prime Minister Andy Bham has said, "It doesn't feel to me that recruitment in the post-pandemic era is becoming fairer. It just does feel harder these days to make that breakthrough, [music] to get that first connection. >> Getting yourself in front of someone is hard partly because of the volume of applications, partly because technology is being used for some of the shortlisting, but there are challenges for employers, too.

Matt Weston is a senior executive at recruitment firm Robert Half. It places over a 100,000 candidates a year globally. It's hard to find the true skills of the individual. I think with AI, you can put in a job spec, it can generate [music] a CV. >> I'm trying to imagine the scene in your office. Are you sometimes sitting there just going goodness more and more and more applications coming in all looking quite the same? >> They are very sy and and that's really really hard um to that's the art of what we do.

The art is justifying or or really going through and seeing who is the right candidate. A recent FT article explained why recruiting is going retro in the age of AI. When every application is polished, polish tells you nothing, one HR executive says, which is where referrals and personal recommendations come in. And as Matt emphasizes, checking out whether the CV actually matches what the candidate can do. >> It all comes down to vetting that very tightly.

Um, a CV is essentially a two-dimensional piece of paper, right? It can have everything on there. Where the art and the skill of where we come in is where we can verify true experience and relevant skills. In other words, even if technology helps, human judgment is still an essential part of the process. >> At this point, it's important to be clear on something that the use of technology to shortlist candidates isn't necessarily seen as a bad thing.

The information commissioner says these tools can enable employers to process high volumes of applications consistently and quickly. Or listen to Kate Young. Her business works with companies which use AI recruitment tools. >> What we do know is that humans making decisions about who to hire is a deeply flawed process and it [music] always has been. I'm sure we can all remember think of a time we were rejected for a job.

We didn't have a firm eye handshake. We didn't make eye contact or some other speurious question. in the interview asked us that [music] actually nothing to do with our capabilities. >> Kate offers a useful reminder that humans haven't always got it right. [music] >> Generally, if the right guard rail rails are present, evaluate on what it's being told to and assuming those things are job [music] relevant criteria, there is the possibility of a better outcome because all that error and noise and bias can potentially be [music] screened out, but only if it's built well.

And you said if the guard rails are there, but that's a big if because there are huge concerns, aren't there, that AI will [music] prioritize or elevate certain types of people. >> Those concerns are very real and is a big if on the guardrails. You're quite right. [music] Um, and the guard rails are many and complicated. Um, and they're everything from who built the system, did they have the right expert [music] expertise to do it, which particular technology they used, which particular AI [music] technique they used. do that well, Kate argues.

And AI can be good for both applicant and employer. Though more broadly, it's not clear how easily employers can adapt [music] their AI tools to meet their specific requirements. And there's another concern. You might think that applying to 10 different companies gives you 10 chances. Not necessarily. If different employers are using the same or very similar algorithms, the same candidates could be favored again and again.

Researchers have a term for this risk, algorithmic monoculture. >> That's not actually a new phenomenon. It has not a lot to do with AI to be honest with you. So if you think back 30, 40 years, we would typically see [music] the same types of people getting the best jobs anyway. You know, those people were well off. They might be [music] members of certain majority groups. they might have access to c certain social privilege.

They might have gone to the best universities [music] and those people would always have had their pick of jobs. And then if you think about assessment before AI, um employers were still able [music] to configure systems to choose certain types of people. >> This is contested territory, both which biases humans may have and which factors may offer reasonable grounds to favor or reject a candidate. And though it's claimed AI can move beyond human biases, given it learns from human data, that is far from guaranteed.

Whatever the merits, if you do get through an AI short list, next may come an AI interview. >> Hi, I'll be your AI interviewer today. My goal is to learn more about you and the experiences, skills, and competencies you might bring to this role. >> Sometimes an AI avatar will ask you questions and then respond to your answers. Other times you're given a set series of questions with each answer recorded as a timer counts down.

The World Economic Forum says AILE interviews can expand opportunities for overlooked candidates, reduce recruiter bias, increase efficiency, and make rejection more constructive. But the experience of these interviews isn't always so positive. >> It's really uncanny. It's really inhumane. And it it's like I'm trying to speak my heart out, be as passionate as I can, but the the persona only sees, oh, are you able to do this?

Are you saying these key words? >> How does it feel when you're talking to a AI generated person to try and get a job? It >> doesn't feel nice. It doesn't feel nice. It feels It feels like I'm putting all this effort in and I I don't see like a like with a normal human, you'd see Yeah. like you know, smiles. got to the stage where I had an an interview where I was recorded, but I was essentially talking to an AI and the AI would basically ask me questions based on what I was saying.

Once I had this AI tell me that I lacked empathy. So that for me was [laughter] a bit was a bit crazy. >> Does it matter to you that there's no evidence of a human being involved? >> Yeah, definitely. It feels like uh I'm never going to be understood by the other side. It feels like I'm talking to a robot. >> I've got to be honest, it feels reasonably bleak that you're not just talking about wanting to get a job. You're talking about wanting to reach the point where you can speak to someone, >> which is even before getting the job, which is even more like, oh, this is such a long process I have to go through to even get my potential seen. >> Question here is how much of this is the fault of new technology?

That is a very real and valid feeling and I have a lot of sympathy with that perspective. Um, of course we all want to meet a human. We want to feel that connection and we want to feel that our time has been valued and our efforts have been valued because applying for a job is emotional and important. I think it's easy to blame AI and AI in hiring has its flaws. But I don't think that's the reason people aren't necessarily reaching a human being. >> Needless to say, recruitment before AI wasn't perfect.

Far from it. But there are some particular challenges with what's happening now. >> So it's more like an arms race now between the the employers and the candidates really because the employers are using AI to sift through applications and the candidates are using AI to better their applications for the role and essentially it will just become AI reading AI. >> They they would suggest um be yourself in your applications like have a a true voice but then at the same time when you actually do that you just get instantly rejected.

AI isn't responsible for every challenge with job hunting. And it may prove very helpful, but clearly there are issues. And as the genie is out of the bottle, it's not a question of if it's used. It's very much about how >> you get behind if you don't use AI today in in this market. It's it's a must unfortunately. >> Yeah. I guess it's a shame, but I guess it's reality as well. >> More like a survival of the fittest now.

Really, it's like you either use AI with your applications, with the keywords, or you don't. And at the end of the day, everyone wants the job. >> Before I finish, I wanted to quickly mentioned that if you're one of those looking for a job, the National Foundation for Educational Research has identified six key skills that you need to stand out from the crowd. And it could be that you already have them. Our education editor, Brwin Jeff, has already looked at this in another BBC News video.

I'll link to that in the description.