Job Seekers Using Genai To Fake Skills And Credentials

Generative AI is being wide utilized by occupation seekers to clone resumes, interviews, and moreover identities — fueling a world emergence successful hiring fraud that’s costing companies millions and making it harder for existent candidates to get hired.

Generative AI (genAI) devices — particularly those for illustration OpenAI’s ChatGPT — are being utilized by galore occupation seekers to enhance, exaggerate, aliases outright fabricate parts of their resumes, screen letters, aliases moreover responses during occupation interviews.

“We’re seeing this a batch pinch our tech hires, and a batch of nan condemnation building and overuse of buzzwords is making it ace obvious,” said Joel Wolfe, president of HiredSupport, a California-based business process outsourcing (BPO) company. HiredSupport has much than 100 firm clients globally, including companies successful nan eCommerce, SaaS, healthcare, and fintech sectors.

Wolfe, who weighed successful connected nan taxable connected LinkedIn, said he’s seeing AI-enhanced resumes “across each roles and positions, but astir evident successful overembellished developer roles.”

Cliff Jurkiewicz, vice president of Global Strategy at Phenom, an HR exertion vendor, said he’s besides seeing a emergence successful fraud during occupation applications and interviews. “It is decidedly thing we person seen for a while, but are seeing a batch much of [now],” Jurkiewicz said.

Candidates are moreover turning to devices for illustration ChatGPT during unrecorded distant aliases recorded video interviews — sometimes reference answers disconnected a 2nd surface fed by an AI assistant. In immoderate cases, it’s not moreover nan existent applicant being interviewed, but personification much qualified opinionated in.

“This is happening globally and crossed industries, pinch nan extremity of getting a much tin personification to walk nan question and reply while personification other shows up for nan job,” Jurkiewicz said. “Clients study that 10% to 30% of interviews — particularly for roles for illustration engineering — impact immoderate level of fraud.”

Workers admit exaggerating their skills

A 2023 study by UK-based StandOut CV, a resume penning service, recovered that 73% of US workers would see utilizing AI to embellish aliases dishonesty connected their resumes — and that was earlier genAI devices had go arsenic communal arsenic they are now. The study besides showed that astir two-thirds of respondents (64.2%) said they’d lied connected their resume astatine slightest once. Conversely, only one-in-five said they’d not been caught lying connected their resumes.

In a abstracted survey by Resume Builder, 45% of respondents said they exaggerated their skills pinch AI devices during nan hiring process: 32% lied astir skills connected their resumes and 30% lied during nan question and reply process.

In general, employers mostly opportunity they don’t person a problem pinch applicants utilizing genAI devices to constitute a resume, arsenic agelong arsenic it accurately represents a candidate’s qualifications and experience. ZipRecruiter, an online employment marketplace, said 67% of 800 employers surveyed reported they are open to candidates utilizing genAI to thief constitute their resumes, screen letters, and applications, according to its Q4 2024 Employer Report.

Companies, however, look a increasing threat from clone occupation seekers utilizing AI to forge IDs, resumes, and question and reply responses. By 2028, a 4th of occupation candidates could beryllium fake, according to Gartner Research. Once hired, impostors tin past bargain data, money, aliases instal ransomware.

For example, North Korean IT workers are reportedly flooding tech companies pinch clone resumes and often usage names that sound American, specified arsenic Mike Smith aliases Thomas Williams. The North Koreans utilize stolen American identities, often obtained done acheronian web marketplaces, and employment VPNs to disguise their existent locations. In immoderate instances, they manipulate convention history files and transportation perchance harmful files to institution systems. These “workers” are typically based successful countries for illustration China and Russia to debar detection.

In January, nan US Department of Justice indicted 5 people progressive successful facilitating this operation. The Justice Department has besides seized $1.5 cardinal and 17 domain names linked to these activities. The FBI and different agencies proceed to show and disrupt specified schemes.

Legitimate IT occupation applicants mightiness beryllium much assured utilizing genAI exertion arsenic AI coding devices go much common. Gartner predicts 70% of developers will usage them by 2027, up from little than 10% successful 2023.

Two sides of nan genAI coin

Emi Chiba, a elder main expert successful nan Gartner Human Resources practice, said genAI makes it easy to create tailored content, truthful galore occupation candidates usage it for resumes and screen letters.

“At this constituent I would presume nan mostly of candidates are submitting applications augmented by AI,” Chiba said. “I’ve spoken to organizations who besides noticed candidates utilizing it during interviews aliases assessments wherever nan campaigner whitethorn publication an reply generated by AI, aliases usage it to thief trade an answer.”

Some candidates usage AI deepfakes — changing their sound aliases quality — to hide their identity, location, aliases truthful personification other tin do nan question and reply for them, Chiba said.

“This happened to maine personally,” Jurkiewicz said. “We hired personification successful Texas who was secretly outsourcing nan activity overseas for a fraction of her pay.”

The counterfeit worker was repeating nan deceptive believe pinch 4 employers astatine once, earning $300,000 to $500,000 a twelvemonth while doing almost nary activity herself, Jurkiewicz said. “It was purely a money grab, and pinch small to nary consequences; there’s nary existent deterrent,” he said. “That’s why we request amended tech to observe this benignant of fraud.”

GenAI, Jurkiewicz said, is some a problem and a solution. It helps surface applicants, and tin besides observe fraud. The institution precocious announced it would beryllium launching a built-in AI supplier for spotting deepfakes astir May aliases June.

Fake candidates make it harder for qualified workers

Another downside to nan increasing flood of AI heavy clone applicants is that it affects “real” occupation applicants’ chances of being hired. “What if you are getting passed complete for jobs by ‘fake’ candidates?” Jurkiewicz said. “There’s an absorbing dilemma.”

Job applicants who clone their credentials will successful each likelihood yet beryllium uncovered and fired, Jurkiewicz said — but by past they’ve accomplished their goal. “It’s a numbers game: group taxable hundreds of applications, hoping immoderate will gaffe done undetected. Even if only a fewer succeed, that’s capable to utilization nan system,” he said.

After moving three, 4 aliases six months, nan clone worker leaves pinch nan salary he aliases she has earned, and moves onto nan adjacent employer, Jurkiewicz said.

The financial effect tin beryllium huge. The US Department of Labor says a bad prosecute tin costs 30% of their first-year salary, pinch immoderate HR firms estimating losses betwixt $240,000 and $850,000 per clone employee.

“Now, ideate nan costs of hiring a clone candidate,” Jurkiewicz said. “They spell done training, do each this orientation. You’re 3 months getting paid, and let’s opportunity you do that crossed nan committee for 100 different employers, because you built this fraudulent network. You’re making a batch of money earlier they observe it.”