Campus Recruiters Grapple with AI

At a recent presentation on AI for campus recruiters and student careers advisors, one in three people said they had not yet used Chat GPT or similar. On one hand that was a surprise given the implications of generative AI continue to dominate discussion around the student/graduate recruitment process. Yet, when it comes to technology there are always innovators, early adopters, the followers and the laggards. So around 30% (the laggards) who hadn’t explored it it shouldn’t be a surprise. But at least they made the effort to learn from the presentation.

What was the key take away? Anything in an application or assessment that is text based or relies on language learning models is at risk of manipulation. From resumes and cover letters, to the text responses from video interviews, even to psychometric assessments.

Despite calls over recent years (mostly from solution vendors) that the resume is dead, it survives. Recruiters and hiring managers like a resume because it provides some factual information about the candidate. Even better if a candidate is able to include context from their experiences that relates to the role or employer.

But with generative AI, the resume (and the cover letter) risks becoming a commodity.

Apart from some factual differences, like university attended, GPA etc., the AI “spin” will do its best to match the student’s content with what the employer wants to read in their application. Look at all these “perfect” candidates!

We could go back to the days when managers would say “I can’t tell one student’s application from another. They all look the same”.

For employers who rely on resumes to shortlist students, it will leave recruiters (or their own AI tools) second guessing themselves.

But the risk of manipulation is not limited to resumes.

Written language assessments are highly susceptible to AI-generated responses.

Psychometric assessment providers are playing catch-up to prevent candidates doing cut-and-paste answers from the likes of Chat GPT. But generative AI will continue to get “smarter”. While candidates can teach the models “here’s my communication tone and writing style” so they can paste personalized responses.

Video interview technology that interprets the language used by applicants is also at risk. We heard from an employer that in their most recent recruitment program more students were “passing” their automated video interview assessment. It seemed odd to them. So they went back to review the student information compared to their video interviews. The results didn’t match up. The suspect? Chat GPT prompting the "correct" answers in the interview.

After all it’s not hard to come up with AI prepared responses to the ten most likely interview questions, customised for each employer. Some recruiters will say "but we can tell if the candidate is reading from a script". True and recruiters might reject every applicant who appears to be reading, whether they are relying on AI or just their own notes.

But on a scarier level, there's emerging technology that counters this. It shows a video interviewee who looks steadily at the camera, when they’re not looking at it at all. The technology makes it appear that way while they read out a generated response to the interview question.

So what does it all mean for high volume student recruiting?

Many of the tools recruiters have used to shortlist candidates will become less effective and less reliable.

Resumes that look alike, other than basics like university and grades. Language assessment results that leave the recruiter with doubt. We may revert to more in-person interviews and assessments in a controlled environment.

There’s no doubt vendors will respond to combat AI with AI. Not all will be successful. While others will develop technology that isn’t vulnerable to text interpretation and language models.

In campus recruitment that technology already exists. GradSift , a platform specifically designed for student recruitment applies AI to factual data for a multi-criteria assessment of applicants. That provides objective evaluations preserving the integrity and fairness of the assessment process, while delivering the substantial efficiencies of AI. Expect to see more developments like GradSift .

#campusrecruiting #graduaterecruitment #aiassessments