Going, going, global: From admin support to legal assistants, schedulers, and sales reps, Solvo helps companies across 50+ industries scale their team with high-quality, vetted global remote professionals.
Move over, “Tell me your biggest weakness,” a new weird interview question has surfaced: “Can you wave?”
Not kidding.
AI-generated deepfakes and candidate fraud are growing in recruiting, and—in response—talent acquisition (TA) pros are learning new ways to suss out the fit of a candidate, and that assessment now must include, “Are you real?”
“Fraud hiring isn’t new…it’s really just evolved,” according to Julia Frament, Head of Global HR at IRONSCALES, a cybersecurity company specializing in email. “Before it was mostly misrepresentation, so candidates stretching titles or fabricating degrees, maybe borrowing experience…maybe ghosting verification, using fake references, using friends to pose as past managers.”
But now the toolbox for candidate fraud is expansive. ChatGPT can write up a perfect résumé and cover letter to match an opening; generative AI tools can create realistic images and content-like responses to interview questions; it’s also relatively easy to design a “real” LinkedIn page for a fake professional.
In fact, up to one-fourth of all job applications could be deepfakes or fraudulent by 2028, according to an April report from Gartner.
“At first glance, it seems like it’s just bad hires, fake resumes, fraud, but it is so much bigger than HR,” Frament said. “It really is about trust and security, which can lead to massive problems within a business.”
System access and security, business data, customer data, financial information, employee information, and business reputation are all at risk, Frament suggested. If nothing else, for the HR or TA pro, dealing with fraudulent candidates adds friction to the hiring process, slows down time to hire, and generates a lot of extra work.
How to spot fraud? Frament said red flags often appear early in application materials.
“The biggest thing that a lot of us are seeing is the overly polished résumés…a strong recruiter is going to see that immediately and raise a red flag.”
Other fraud indicators include a noticeable mismatch between verbal and written communication, an unusual cadence on camera, or an audio delay that’s a little off, she said.
“I work for a 100% remote company, and I definitely think we’re more susceptible because it’s easier for someone to fake their location and identity when they’re interacting with us via Zoom versus having to show up for an in-person interview,” Bonnie Dilber said, a recruiter at the workflow automation software platform Zapier and content creator who posts to social media about recruiting. “I will say that I know someone who hired someone who was lying about their identity, and that was for an in-person role with a fully in-person hiring process.”
Dilber said there are obviously bad actors out there looking for an in with a company to target proprietary, financial, sales or personnel data, but there are other situations where AI engineers are engaging in this process to build tools to “hack interviews” and sell a product.
“They’ve done it because they want to prove either a proof-of-concept for their tool, or because they need to figure out what’s going on in the interview process for that company so that they can design something they can sell,” she said.
But bad actors aren’t the only candidates who are using generative AI tools to polish (or heavily write) their application documents, Dilber added.
“We get huge volumes of applications, a lot of the content in those is AI generated, and you don’t actually know if it’s AI generated because it’s a fake, fraudulent person, or if it’s just someone using AI to speed up their application process, both of those things are happening,” she said.
Dilber also noted that any flagged issues do not automatically mean the candidate is fraudulent, but raises that possibility as the process proceeds. Her team checks that IP addresses match locations where candidates purport to be, does a double take if virtual phone numbers like Google Voice are used, and notes VPN use.
Frament told HR Brew that IRONSCALES, too, developed a set protocol for dealing with potential fraud in the hiring process: “dig in quickly, then move on.”
What’s HR to do? Frament suggested normalizing awkward conversations about the 800-pound fake gorilla in the room. If you’re unsure about the authenticity of the “person” sitting in front of you, ask them off-script questions, if they’re having connectivity issues that might explain any lags or glitches, or ask them to wave their hand in front of their face and look for video glitches, she said.
“It sounds silly right now because it’s not something we’re used to, but I do believe that it’s going to be…normal,” Frament said. “Little things like watching people’s movement, really being mindful of any glitching…and taking note of that if you have any concerns.”
The gut feeling isn’t always correct, Frament said, suggesting that candidates each have unique individual personalities and sometimes it’s just a “weird” vibe. She also pointed out that in international hiring, language and cultural barriers can also add to the challenge since many US norms are not shared abroad and vice versa, so sometimes raised flags don’t indicate fraud.
“We have always been trained to identify skill and cultural fit,” she said. “Now we’re going to have to be trained to identify authenticity and verification.”