Job Search 2.0: Artificial Intelligence Impact in Candidate Pre-Screening

Room 42 is where practitioners and academics meet to share knowledge about breaking research. In this episode, Huiling Ding explains the impact of AI tools on job application materials and strategies to address applicant tracking systems, resume screeners, and on-demand video interviews

Season 1 Episode 17 | 47 min

Subscribe: Listen on Apple podcasts  Listen on Google podcasts  Listen on Spotify  Listen on Stitcher Listen on Amazon music Listen on iHeartRadio Listen on Castos

Transcript (Expand to View)

[00:00:13.120] – Liz Fraley

Good morning everyone, and welcome to Room 42. I’m Liz Fraley from Single-Sourcing Solutions. I’m your moderator. This is Janice Summers our interviewer, and welcome to Huiling Ding. Today’s guest in Room 42. Professor Ding teaches technical communication at North Carolina State University. Her research focuses on Intercultural Professional Communication, Technical Communication, Risk Communication and Epidemic Communication. Her recent projects have been exploring the connection between artificial intelligence, communication technologies, risk communication and social justice as the principal investigator of a large interdisciplinary, National Science Foundation C-Accel Grant, she’s been leading her team to examine how AI tools have been transforming the job market and job screening processes in the US. Today, she’s here to help us start answering the question, What Assumptions about Pre-hire Screening Criteria and Procedures are in use, and What’s their Impact on Job Application Materials and Strategies? Welcome.

[00:01:17.210] – Huiling Ding

Thank you. That was a great introduction.

[00:01:20.430] – Janice Summers

Huiling, we are so excited to have you here and there is so much to talk about, like your interest in your research, in your background, in your teaching is so fascinating. But one of the things I want to ask is, you know, the topic on point of AI and the recruitment and job search process, how did you get involved with this particular– you had to write a grant for this, right? This is part of your grant research. Yeah.

[00:01:50.180] – Huiling Ding

Right, right.

[00:01:50.180] – Janice Summers

What’s your origin to this, What inspired you?

[00:01:55.040] – Huiling Ding

Great question. Thank you Janice, so my interest started, I think three years ago, I watched very closely the AI competition between the US and China and how it was kind of like an arms race, right, okay who will be the next international leader in this emerging field that will have great potential in terms of technological breakthrough. So I started with some, intercultural comparison of AI policies in the US and in China. And then AI started to be more widely used in communications, I started to see, you know, like user testing for AI associate writing. And then some of my students would come to me and say, hey, Dr. Ding, I can’t get past the video interviews. I’m like, what video interviews? And they were like, I have you know, I’ve caught some recruiters say, OK, we like your resumé, right, we’ll invite you to a online video interview and then you’ll be talking to a camera and then that will be screened. And they were like, I never hear back from them after that video interview. So that got me interested, and I did some quick search to make sure I won’t be doing my students a disservice when I teach them how to write, right, for job search.

[00:03:31.160] – Janice Summers

Right.

[00:03:32.110] – Huiling Ding

Right, and then I realized, oh, oh my gosh, those are not really like the traditional video interviews. Those are video interviews that will be evaluated with image recognition, facial recognition, voice recognition, AI tools. I was like, oh my gosh, what’s going on here? And I looked for–

[00:03:52.700] – Janice Summers

For something that seemed a little innocent.

[00:03:55.340] – Huiling Ding

Yeah.

[00:03:55.850] – Janice Summers

You could, you know, you unpack all those layers and then it’s a lot more complicated.

[00:04:00.320] – Huiling Ding

Right, right, and you, once you look beyond that, it’s like, OK, who is doing the evaluation? How are those videos evaluated? How are candidates ranked? Especially when I found out, you know, those who have more difficulties going beyond video interviews tend to be, you know, ethnic minority students or non-traditional students. And that was really like a red flag to me. So I looked into that and then found out there were already all kinds of the so-called HR technologies or automation of recruiting or AI assisted recruiting. So that was also the time when one of my friends passed their request for proposal of P-form National Science Foundation to me, which deals with AI and future jobs. So it was perfect timing, and I managed to put a team together working on, you know, retraining reemployment assistance for manufacturing workforce and, well, other co P.I, my co Principal Investigators worked on the engineering and AI side. I focused mostly on AI and recruiting and employment. So basically, how do we get screened today in the job market?

[00:05:30.770] – Janice Summers

Wow, and there’s just so many challenges with relying on AI, from a recruiting perspective, because it’s only as good as who programmed and how they programmed, because it’s trying to interpret everything right, it’s trying to interpret a lot.

[00:05:50.970] – Huiling Ding

Right.

[00:05:52.860] – Janice Summers

How do they account for all of the facets of a human being, it’s–

[00:06:01.430] – Huiling Ding

It’s difficult, so–

[00:06:03.650] – Janice Summers

Yeah.

[00:06:04.190] – Huiling Ding

Yeah, so for instance, I was trying to map the workflow of automatic recruiting just in terms of resumé screening, so we heard the use of applicant tracking system.

[00:06:20.630] – Janice Summers

Right.

[00:06:21.470] – Huiling Ding

Job ad passing, resumé passing and then job ad and resumé matching and then, ranking of candidates. So it’s the collaborative efforts of different AI tools, especially natural language processing tools. For instance, my team read a lot of their engineering AI publications, natural language process and publications early in maybe before 2010, before the tools became commercialized, and those papers talk about possible ways to design those tools from engineering perspective. So it’s really interesting to see how the job, any job ad will be broken down into maybe 80, 100 data points. So like, let’s say it’s a 300 words job ad it will be automatically broken down into 80 data points with some of the ones being like, OK, top cut, right, if you don’t pass this, your resumé will be rejected right away. And you can imagine those data points would be location, years of employment.

[00:07:41.540] – Huiling Ding

Those are like verifying, so I heard about how candidates were rejected because they were one month short of their work experience requirement and AI you know, they don’t negotiate with any human beings, right, if their job ad asked for three years or 36 months and I have only 35 months, most likely I will end up being in the rejected pile, right, and the numbers show that especially for the larger companies like fortune 500 companies many people openly will have maybe 250 and more applications and Janice, when you talk about recruiting, I looked into recruiting a little bit there, the staff inside, right, and I realize how their staff inside also have to compete, right, for that one job opening because they will be paid by, you know, how many positions you’ll help to fill. So there is also– not only that employers need to fill the jobs much more quickly to reduce the cost of a transition from one person to another, recruiters have to run very quickly to make sure they grab the right candidate so they can get paid for that one job field. So that was really interesting.

[00:09:05.990] – Huiling Ding

So making the cut right, and then there were additional data points such as degrees, right, those are all pretty quantifiable measurable things. Degree, okay, we need a degree in Techni communication or professional writing or writing composition or journalism. That’s easy. However, when we move on to the skills part, it’s messy. Job titles are messy, right.

[00:09:35.670] – Janice Summers

Oh my God, yeah job titles are, especially I think in this field, I think especially in this field,  job titles are a mess. They’re a mess, because it just seems like everybody keeps changing their job titles.

[00:09:51.940] – Huiling Ding

Right, right.

[00:09:53.000] – Janice Summers

For some reason, and if you look at like the skills that are required in the job that they do, you know, doesn’t change all that much with the job title.

[00:10:06.650] – Huiling Ding

Yeah.

[00:10:07.730] – Janice Summers

Changes. So I don’t know how you even go up against that.

[00:10:11.690] – Huiling Ding

So we look into that as well, and the better ways to do it is, you know, so the computer scientists we work with, they were trying to look for a range of job titles like relating terms. And if we in techcomm and UX writer or content strategist or you know, related job titles should be included, right, but depending on which applicant tracking system or job matching system is being used, some of them may be better at doing that than others. So… Right, right, and that’s a great question. I think, you know, you can describe your job responsibilities, remember, job titles is one way of screening and then the most important thing will be the skills the most– the toughest to measure, right, so, for instance, I was really fortunate to also have access to some labor market analytical tools such as Burning Glass, and I can talk about that very briefly as well.

[00:11:29.750] – Janice Summers

Yeah.

[00:11:29.750] – Huiling Ding

Yeah, if you pay attention to how labor market analytics are compiled, the labor of– the Department of Labor, offer O*NET, right,  which is more static, which is updated once every two years.

[00:11:51.130] – Janice Summers

Right.

[00:11:51.910] – Huiling Ding

If you look for O*NET and technical writing, you can see the type of skills and job titles and payment salary and so on right in that.

[00:12:03.670] – Janice Summers

How did their job titles, because they lag the industry right, how do their job titles match what commercially is going on?

[00:12:11.860] – Huiling Ding

They don’t.

[00:12:12.660] – Janice Summers

They don’t.

[00:12:14.420] – Huiling Ding

All that and they don’t.

[00:12:17.570] – Janice Summers

Yeah.

[00:12:18.410] – Liz Fraley

So how bad is it, like what do you see?

[00:12:22.660] – Huiling Ding

O*NET they don’t, they are so bad that if you look at a zip code, right, it’s how we as a field are matched with academic programs and I as a graduate program administrator, really hate the system because the zip code, if you look at the zip code, right, so it’s like, OK, we don’t have graduate students graduating from our programs or very few right, like in the past 10 years we may have 15 students graduating. So it’s not measuring the job market very well. So what labor analytics tool such as Burning Glass, what it does is to come up with its own professional codes, so it’d come up with a much wider range of job titles for the type of jobs that Techni communicators may have. So that helped to catch that a bit better. Okay, so that’s a job title and then the skills, so my engineering team, when they measure the skills, what they use is what they call skill vectors. And that’s quite commonly used in AI as well. So they come up with the categories of skills that a job requires, right. So let’s say if we are in Techni communication. The categories we have will be writing, editing, right, and then specialized tools and then maybe usability, you know, research and so on, so forth, right, and then they use synonyms. So that the algorithm the NLP tools will be looking for synonyms and they try to define those synonyms so that the algorithms will recognize related terms, right.

[00:14:34.490] – Huiling Ding

So that’s how their skills are defined in their other occupation level, occupation level. And then when we move on to individual jobs, different jobs require different things. So if we want to make sure we can be matched, our resumé can be match-able through the one job ad, we’d better do some, you know, serious manual matching to make sure, Okay, here are the skills that are required, and here’s how I reflect and prove I have those skills in my resumé. So we want to make sure our resumés are match-able.

[00:15:18.170] – Janice Summers

Right, so you have to do your data analytics and start analyzing all the various career paths as well to make sure that, you know, you’re guiding your career on the right trajectory, especially when you’re starting out.

[00:15:33.080] – Huiling Ding

Yes, yes. Definitely.

[00:15:37.400] – Janice Summers

Go on I’m sorry I didn’t mean to interrupt you.

[00:15:40.340] – Huiling Ding

No no no. So, we talk about data points right, so a job ad will be broken down into over 80 different data points, and then my resumé, one-page resumé will be broken down again into 80, 75, 80 different data points, location, okay, work history, how long have you been in this position, how long have you been in this field,  and then educational background, if they say, Okay, a Masters Degree is preferred and I have a Bachelors Degree, most likely if I have a competitive pool of candidates, I won’t get anything, right. So those are going to be matched, well, okay, it’s almost like you create a table, I think you can probably manually break down the job ad into those important data points and then {Inaudible}, okay, do I match? How much do I match? Right, I think that’s definitely doable and that’s exactly what the AI tools are doing on a much larger scale very quickly, every resumé gets maybe 2 seconds in the matching process.

[00:16:49.490] – Janice Summers

Well, but I mean, it’s kind of a necessary evil. I mean, it really– because when you get flooded with resumés, you’re estimating 250 resumés is very small. I don’t think that’s really what recruiting– I can tell you from a recruiters perspective, that’s kind of like a small pool, that’s a tiny pool. That’s almost enough for me to personally have a conversation with all those people so they get volumes and volumes of resumés and so they have to read through it somehow. So, those are pretty understandable, right, so far when you’re talking about the resumé matching to the job description. So what are some of the other AI techniques that people use in screening candidates out?

[00:17:44.860] – Huiling Ding

Right, so for the job ad resumé matching, 75 percent of candidates will be screened out.

[00:17:53.560] – Liz Fraley

Wow.

[00:17:55.480] – Huiling Ding

 Yeah, 75 percent, and then depending on the recruiting agency or staffing agency, and depending on the employer, you know, maybe the rest of the 25 percent, maybe invited to have automated media interviews and then that will be screened as a second round. So eventually, the user journey or the life circle of automated candidate screening will end up with a shortlist of three to five candidates which will be passed on to the employer. As you know, here are your candidates that are worth your time of interviewing, engaging ways and eventually given an offer to, right. So in that process, we also have video interviews that will be done at our own home, right, and will be recorded at our convenience at any time, right, and then will be collected and automatically evaluated and ranked by the recruiting agency. And as I said, those will be evaluated with facial recognition and voice recognition tools. So it was really interesting to say okay, some of the early tools right, they will be looking at, you know, eye contact and facial expressions and then they’ll be measuring intonation, change of tones, right, whether you are engaging with the camera, right, during the video interviews to show your enthusiasm.

[00:19:50.690] – Huiling Ding

And very few actually look at the actual content of what I as a candidate talk about. So that was–

[00:19:58.250] – Janice Summers

You could be talking about anything, it’s just interpreting are you telling the truth? Right, I mean it’s looking at all these as a– I just don’t, I don’t know. To me personally, I think the video interviewing is a pretty dicey area because I see so many potential for flaws, I mean how do you account for people who have multiple languages that they speak, how do you account for people who have disabilities, how do you account for, you know, a lot of different face structures, I just don’t know. What are your findings on this whole video conundrum?

[00:20:40.390] – Huiling Ding

So the video interviews I think we definitely need, you know, access to better cameras, you know, like lighting background, right, and then those of us who have darker skin colors may have a tougher time because of the way their algorithms were trained right, with mostly Caucasian users, face images, and then there were also, I don’t know, intonation, I mean, how do you charge the enthusiasms based on intonation, I think there’s a lot of biases built in those algorithms.

[00:21:26.410] – Liz Fraley

Yeah.

[00:21:26.410] – Huiling Ding

Right, so it’s so problematic that Illinois, the state of Illinois, actually issued regulations saying, you know, we don’t allow the use of video interviews, I think back in 2018 and then there’s also a requirement of being transparent, what do you measure,  how do you measure it, right, depending on what kind of institution we’re talking about, our employers we are talking about. Another thing that’s widely used is social profiling. So, I was shocked to hear that social profiling, so basically social media content will be monitored even in college admission processes for teenagers, right, who are applying for admission into top 10 let’s say Ivy League Universities, they’d better monitor their social media counterparts very closely. And for us professionals, you know, there are AI tools that will be screening the web, looking for any online activities content. Even, you know, the type of content we click like or we comment on, in LinkedIn or Facebook to come up with, you know, like some analysis of the applicant’s personality.

[00:22:57.390] – Janice Summers

Exactly, and they track that to make decisions on how you will behave as an employee, right?

[00:23:06.680] – Huiling Ding

Right.

[00:23:08.970] – Janice Summers

Everything you do digitally has a footprint.

[00:23:12.880] – Huiling Ding

Right, and we have to– I don’t know, it’s kind of like the message, what is the message we give to our students? I think I talk about this in my graduate class, the Capstone class, and students asked, is it in my interest to post less or should I still be visible in social media? And I was like, I don’t know, I think if you are comfortable sharing your personal, like, daily activities in Facebook, go ahead, but stay professional, right, don’t be too personal, you know, think about it and anything that’s published online will be there forever. Even though you delete the content, it may still be found-able if someone else sees that, right, so be extremely careful.

[00:24:10.310] – Janice Summers

Yeah, because what you do can live with you forever.

[00:24:14.720] – Huiling Ding

Right.

[00:24:15.350] – Janice Summers

You might always have to be trying to explain something, but when it comes to employment, you might be screened out as a result of some of these findings yeah.

[00:24:27.170] – Huiling Ding

Yes.

[00:24:31.510] – Janice Summers

And I think that’s, you know, that’s a conundrum in the recruitment processes, because they have to screen out so many rather than filter people in.

[00:24:40.300] – Huiling Ding

Yeah, right, the message I found to be really interesting is, you know, H.R. says no and hiring manager says yes. So the first round of screening, the AI will reject 75 percent and then the rest of the 25 percent will be further screened until you know, if I am lucky enough, I’d come there while their last few finalists and then I will be in contact with any human being in the process.

[00:25:17.090] – Liz Fraley

It’s so strange to even think about.

[00:25:20.270] – Huiling Ding

Right.

[00:25:21.850] – Janice Summers

Yeah, I mean, and I get the first level, I get the resumé right. I get that and I get the automated tools that will kind of help match things and filter, right, because you’ve got to do something, it’s a lot of manual labor to try and go through that, especially when you’re faced with, I tell you more than 250–way more than 250 candidates, especially if it’s for a prime job, you know, or a popular job. You’re going to get a lot more than that, so you’ve got to do something to screen. But–

[00:25:54.970] – Liz Fraley

It’s got to be even harder when you don’t have any good keywords matching because both sides are changing all the time.

[00:26:03.920] – Janice Summers

Yeah.

[00:26:06.400] – Huiling Ding

Yeah.

[00:26:06.760] – Liz Fraley

Right, people chasing titles and employers changing the job to match whatever title they have and think it is too, as well as the unique skill sets for the position they’re hunting for.

[00:26:21.730] – Janice Summers

That’s true, and I think that’s why you can always, like, look for and do your research right, it becomes like if you’re looking for a job, it becomes first and foremost, you’re a researcher.

[00:26:35.700] – Huiling Ding

Right.

[00:26:37.900] – Janice Summers

Right. You have to go out there and do your research and look not just at one company that you want to go to, but a bunch of companies that are looking for this type of field. And this brings up a really good point. The title thing is just a nightmare, you know, and that bucket of titles, I don’t even know, It’s like huge now and it changes constantly. So but it’s the skill sets and you have to look at what are they looking forward to meet, and then I think that becomes pretty easy or, well, less complicated to do a data query and then compile the information and then look at your own experience and figure out do you have it or not. But the video and that, how do you go up against that?

[00:27:31.650] – Huiling Ding

I think–

[00:27:31.980] – Janice Summers

Wear a mask? I don’t know, have somebody else sit in for you? What do you do? I mean.

[00:27:39.750] – Liz Fraley

Yeah, right.

[00:27:40.980] – Huiling Ding

I think the algorithms will have its own screening criteria, and it’s never going to be transparent. So we–

[00:27:52.590] – Janice Summers

That’s the problem.

[00:27:54.030] – Huiling Ding

Right, right, so the only thing I think we can do is to ask the question, what are the commonly used criteria right, in a video interview evaluation? So and then how can I do my best to present myself right in that interview, knowing those will be the criteria that will be used to screen candidates. So, I mean, other things like how I look right in terms of age or skin color or my accent, there’s nothing I can do about it that’s me right. And other things, I mean, there are things that will be under my control and there are things that will be out of my control. So I think my suggestion for students will be, you know, do your best to be literate, you know, knowing what the truth is all about, how it will be evaluating who you are and then be yourself and do your best.

[00:28:58.400] – Huiling Ding

We can’t hire someone to interview for us, there’s no way that will work. So there are limitations, you know, how much we can do, right, how much control we have, I think there’s always limitation there and it’s just my life, right? If we have a in-person interview, if one of the committee members happen to, you know, look for a male candidate and I’m a female candidate, I’m already at a disadvantage, right, those are they’re all subjective expectations and biases in the human process interviews as well.

[00:29:38.460] – Janice Summers

Well, the hiring, I mean, ultimately is subjective right?

[00:29:42.780] – Huiling Ding

Right.

[00:29:43.190] – Janice Summers

Right, and that’s really I think what you’re trying to do is run through the gauntlet to get to a very subjective process, right. So transparency and AI is one thing that should be out there that’s not there now. I don’t know that the video is going to go away, do you see that growing or shrinking?

[00:30:10.270] – Huiling Ding

I think it’s growing, like the applicant tracking system has become increasingly more popular just because of their costs, cost of hiring, the cost of waiting to fill one position, right, and then the video interviews, it’s the same thing. If I can interview 20 candidates in one day without using my own schedule, right, that’s saving me culminating days of working of my week. So it’s definitely cost effective.

[00:30:44.020] – Janice Summers

Yes, and I think that’s a huge factor in pushing that argument and in marketing those type of tools.

[00:30:51.190] – Janice Summers

Yeah.

[00:30:51.890] – Liz Fraley

Yeah.

[00:30:53.520] – Janice Summers

So the state of Illinois case, you said, came out with– they banned, could you expand on that, did you say that they banned you as a pre-screen? You mentioned earlier, the state of Illinois said something like–

[00:31:13.700] – Huiling Ding

Right, so let’s see, I think, I can’t remember very clearly, I think it’s a regulation saying we don’t allow video interviews, automated video interviews to be used in the state of Illinois, whether it’s to provide discrimination or it’s because of the ethical concerns they say. I definitely need to go back and read a little more closely. That’s the only state that have issued any regulations about video interviews.

[00:32:01.420] – Liz Fraley

Fascinating.

[00:32:05.890] – Janice Summers

So, how would they– because this is going to grow, right, video you’re seeing that as a growing trend in the recruitment process. And it’s growing at the university level as well for admittance into university the video?

[00:32:23.220] – Huiling Ding

Not yet.

[00:32:25.060] – Janice Summers

Not yet, but the social part is the social media.

[00:32:29.510] – Huiling Ding

Yes.

[00:32:30.460] – Janice Summers

Yeah.

[00:32:32.710] – Huiling Ding

So social profiling is definitely there, usually, I think when we hire, how many of us will go online and do a quick search of our top candidates to say, okay, what does she actually what is he or she doing right, we don’t have to do that manually, some of the applicant tracking system can do that very quickly and then present, one paragraph of summary about all candidates. Just one paragraph.

[00:33:10.810] – Janice Summers

Yeah.

[00:33:11.470] – Huiling Ding

And what they look like online, right, and what their activities look like.

[00:33:17.850] – Janice Summers

The types of things they like.

[00:33:20.790] – Janice Summers

I don’t know how much detail we have in those paragraphs of description, that definition, they will summarize what’s going on in this–

[00:33:30.480] – Janice Summers

Just scrolling on a page and staying on a page along to all this data is being recorded by somebody and something’s going to happen with that data. So any time you’re on a digital and the funny thing is if you’re not on digital presence that sends up another flag

[00:33:42.600] – Liz Fraley

You’re not social yeah

[00:33:46.130] – Janice Summers

That also shows your complete avoidance of social media will show.

[00:33:51.870] – Liz Fraley

And the professional sites.

[00:33:54.560] – Janice Summers

Yeah, yeah, send the ball another signal.

[00:33:57.820] – Huiling Ding

Yes.

[00:33:59.010] – Janice Summers

So, with all of this AI, it’s amazing anybody gets hired.

[00:34:09.447] – Huiling Ding

Student’s still get hired.

[00:34:09.870] – Janice Summers

I know it does happen, it does happen, but I think it’s the ethical concern that I have for people with disabilities and people who don’t fit with that software engineer who is writing algorithm in the darkroom, locked in their own sphere of–

[00:34:30.650] – Liz Fraley

Describing himself

[00:34:31.740] – Janice Summers

Reality. You write or you know, like that’s how these things are being written by a very limited pool.

[00:34:40.460] – Liz Fraley

Let’s test this, hey, buddy, make a video for me.

[00:34:43.160] – Janice Summers

Yeah, video you.

[00:34:45.590] – Liz Fraley

Yeah.

[00:34:46.180] – Janice Summers

Yeah, that’s a good sampling of the population, I just worry about those things. And so what do I do if I don’t fit that mode. What do I do, if I have a speech problem, if I have a stuttering problem or if I have a thick accent,  if I am not a very, you know, overly confident person, how do I meet that challenge of having to be videoed.

[00:35:18.980] – Huiling Ding

For video interviews?

[00:35:20.510] – Janice Summers

Yeah, I think the thing is, you know, do your best, get prepared, and then there are plenty of positions today, like that still don’t require automated video interview. So maybe go for those as well, depending on which industry you are in, right, I am an Asian graduate student applying for jobs in business versus in I.T, if I get asked to use video interviews, I may do better in the I.T field than the business field, right. It depends on, you know, what industry, the employees in the industry look like in terms of their ethnic composition. And then there are still plenty of jobs that don’t require automated job interviews, so I think the students I talked about, they eventually got hired through in-person connections, moving from Co-op or intern positions to full-time positions because their colleagues and their employers saw their, you know strength, right.

[00:36:44.660] – Janice Summers

Yeah.

[00:36:45.440] – Huiling Ding

And their– the sense that the value added, the sense that they can contribute to the organization as a whole. So I think there are alternative ways to get there.

[00:36:57.200] – Janice Summers

Right, yeah, I don’t think that that has ever changed since the beginning of time, since employment first started back, whenever it first started. It really is a relationship, and those are the strongest employment factors it’s who do you know, right, the number one thing the recruitment is always asking people who do you know, right, because you want to build that pool, that resource of people. And when you’re bringing people into companies a lot of times it’s who do you know, unless you’re intentionally trying to diversify your workplace and you want to know who they don’t know right, because just hiring people who know you is not going to end up with a very rich, diverse workforce. It’s going to end up with a very slanted workforce. So, but relationships do matter and they do influence.

[00:37:53.400] – Huiling Ding

Right.

[00:38:01.280] – Janice Summers

Yeah, so some of the things to kind of help out with the video assessment, maybe do, just be yourself, and let it go and don’t stress about it so much because there are other employers. The other thing is we don’t know or you don’t know when you’re video interviewing what factors they’re measuring, right, so I think your advice for them to be confident and just be themselves is I think the only thing you can do right now when it comes to video.

[00:38:35.560] – Huiling Ding

Practice, practice.

[00:38:37.210] – Janice Summers

Yeah, practice, practice, practice and not stress about it because it’s not the end of the world and it’s not the only employer.

[00:38:47.470] – Huiling Ding

Right. Well, the signs I think will be helpful is to, you know the criteria, right, so you can do some video recording of your own interview. Like if it’s a three minute interview, we can find commonly asked questions and video record ourselves talking to camera and then look at our performance and say, okay, here are some of the things I tend to do which will not be helpful, right, I think we can make strategic adjustment and changes to help us to stand out better in those automated processes.

[00:39:27.640] – Janice Summers

And now are there– are you finding that companies are knowledgeable of the fact that this could adversely impact somebody with a disability? Are they putting that as a front before the video requirement?

[00:39:46.390] – Huiling Ding

That’s a good question. I’m not sure you know, we do know that people, for instance, employees, candidates with autism, like they won’t feel comfortable engaging with anyone, right.

[00:40:05.850] – Liz Fraley

Right.

[00:40:06.570] – Huiling Ding

In terms of eye contact, so they will have difficulty getting through the video interviews and when competing with other candidates. So, I don’t know how companies, employers come at the consideration of disability and diversity in their screening process. Those criteria will be there, and unless they specifically call attention to the need to, for instance, to bring more diverse candidates, people with disabilities right into their finalists, that won’t happen in the automation process.

[00:40:52.010] – Liz Fraley

Can someone ask, I don’t know how you get the answer, but can you ask about what AI tools someone is using. Right, as a candidate, can say, hey, which tool are you using to screen? Right, and so you can sort of point out well this tool puts emphasis on these things, this one puts emphasis on those things, or is there a way to get it that even?

[00:41:18.280] – Huiling Ding

 Yeah, one of my graduate students tried to do that while doing some back-end coding analysis, and most of the responses she got was no response, no one responded. And when a couple of their companies responded, they would say it’s our proprietary codes and we can’t share that with you. So I think there is definitely the black box.

[00:41:49.750] – Liz Fraley

Can an employer ask about the tool being used by recruiting companies so that they can know where it’s plus or minuses are?

[00:41:59.230] – Huiling Ding

Yes, you will get a lot of marketing materials you know, like and choose the tracking system yeah

[00:42:08.260] – Janice Summers

Very carefully scripted marketing–

[00:42:10.980] – Liz Fraley

Yes, yes.

[00:42:13.660] – Huiling Ding

You’ll have so many vendors to choose from. So of course it’s technical marketing tools, right, and those materials aren’t technical marketing documents, but they offer more insights than what’s publicly available about how the tools are viewed, how they screen candidates.

[00:42:39.310] – Janice Summers

And, you know, unless you’re on the inside in the H.R. department and an employer that’s using video, you don’t know how they’re weighting the video, right. The video might not be the determining factor, it’s just weighted. So it’s giving a certain grade, but it’s not the only factor that they use. So that’s probably why also they’re careful in couching it. And I think your advice to be as comfortable as you can and own yourself when you’re videoing. And if you’re a person who’s nervous about videoing, then video often, right, you know, honestly, be who you are because who you are is how you are when you show up at work anyways. So, and there is the factor of, you know, there’s more than one job out there, right, and not all of them are going to be a fit for everybody. So if they don’t select you, they’re probably not right for you anyways, maybe not right now.

[00:43:45.120] – Huiling Ding

Right.

[00:43:45.690] – Janice Summers

Maybe in the future, so don’t– it’s not a personal insult to you because you will get a job, you will find a job. It’s just you need to go in and look at other places. I think the most important thing is that becoming a data analyst you were talking about earlier, about analyzing the data and researching the field and staying engaged in the field that you’re interested in, because those relationships that you make there are going to matter.

[00:44:17.470] – Huiling Ding

Right, right.

[00:44:18.750] – Janice Summers

And it’s like getting to know a good recruiter, a good headhunter. Just understand, they’re not your therapist.

[00:44:25.530] – Huiling Ding

Right.

[00:44:27.660] – Janice Summers

They are not. And they can only do what they can do by getting to know good recruiters and having a good relationship with good recruiters, because that’s what recruiters do, is they’re always engaged in what’s happening and what employers are trying to do. And they touch more than one employer.

[00:44:43.680] – Huiling Ding

Right.

[00:44:44.340] – Janice Summers

So, having a good relationship with a good recruiter, I think is key.

[00:44:50.380] – Huiling Ding

Right, so one of the exercises I would encourage all my students to use before they write their resumés is to use Corpus analysis tool called Concordance or Antconc, so they can allow students to maybe select the top 10 job ads they are interested in applying for using a specific title and then compile those job ads into one document and save it as a plain text TXT file, and then Antconc will process through those job ads and come up with a list with frequencies for the keywords so they can go through those keywords and look at co-location words, what words go together, and that will give them some really important insights into, you know, their important skills emphasized by these job title, by these type of jobs in general. They can try to use the language and the skill, highlighted those skills when they apply for jobs. And then when they apply for individual jobs, of course we want to tailor our application packages to make sure we reflect that what employers need as well. So, Janice, as you said, it’s a lot of you know,  analysis, decompositions. So I want to emphasize computational thinking, right, remember, job ads are broken down into 80 data points and our resumés will be, you know reduced to data points as well, so how do you decompose job ads, how do you decompose your own resume into those data points and make sure you can do as much matching as possible? I think that’s all the homework anyone can do with a simple tool such Antconc.

[00:46:52.390] – Janice Summers

Yep, exactly.

[00:46:54.330] – Liz Fraley

And here we are out of time. What a great way to end it. Some advice on what to do and how to do it.

[00:47:00.590] – Janice Summers

Yes, it’s been such a delight talking to you. I can’t believe our time’s up, we’re actually over time. I was ignoring Liz telling me time is up, I’m like, no.

In this episode

Professor Huiling Ding teaches technical communication at North Carolina State University. She is the Director of the MS in Technical Communication program and a University Faculty Scholar. Her research focuses on intercultural professional communication, technical communication, risk communication, and epidemic communication. Her recent projects have been exploring the connections between artificial intelligence, communication technologies, risk communication, and social justice. As the principal investigator of a large multidisciplinary NSF C Accel grant, she has been leading her team to examine how AI tools have been transforming the job market and job screening processes in the U.S.

The landscape of searching for work continues to be radically and permanently changed. Ever increasing demands for recruiters to present the exact ideal candidates for job openings has created more reliance on Artificial Intelligence (AI) for pre-screening and matching volumes of candidates against defined criteria.

To understand the impacts of emerging AI-augmented pre-hire assessment tools, this talk will examine assessment tools such as applicant tracking systems, resume screeners, and on-demand video interviews. We will also examine assumptions about pre-hire screening criteria and procedures used by technology and the impact of such tools on job application materials and strategies.

Resources

Email: hding@ncsu.edu

Website: https://huilingding.wordpress.ncsu.edu

LinkedIN: https://www.linkedin.com/in/huilingding

NSF grant website: https://werise.wordpress.ncsu.edu/aitools

Rhetoric of a Global Epidemic: Transcultural Communication by Huiling Ding

Resources mentioned during the show

The Illinois statute about video interviews: https://www.shrm.org/resourcesandtools/legal-and-compliance/state-and-local-updates/pages/illinois-artificial-intelligence-video-interview-act.aspx

AntConc for corpus analysis: https://www.laurenceanthony.net/software/antconc/

The CIP code that focuses on instructional programs: https://nces.ed.gov/ipeds/cipcode/Default.aspx?y=55

The information about technical communication is listed here: https://nces.ed.gov/ipeds/cipcode/cipdetail.aspx?y=55&cipid=89226

The O*NET information about technical writers as a bright-outlook profession: https://www.onetonline.org/link/summary/27-3042.00


Support TC Camp

Everyone at TC Camp works hard to produce educational public service projects. We hope you’ll consider supporting our efforts to help us keep the innovative programing going.

Note: When you purchase something after clicking the links to Amazon on this page, we may earn a small commission. Read our affiliate link policy for more details.

Leave a Reply

Your email address will not be published. Required fields are marked *

Scroll to Top