Lifting the lid on candidate fraud

Many professionals remain unaware of the threat posed by AI-generated job applications despite their far-reaching implications, experts have warned.

Key to combating candidate fraud is understanding that it extends far beyond small omissions – or additions – to résumés, according to WorkPro founder and CEO Tania Evans.

While the phenomenon originally presented as falsified references, inflated employment and salary history, the introduction of AI means that what once simply posed a risk to a candidate’s credibility can now compromise the hiring process.

Kingston Reid partner Duncan Fletcher added that such fraud can risk organisational security because it “covers everything from lying about a qualification or capacity to perform a role to the criminal act of stealing another person’s identity in order to gain employment in a particular role”.

In this way, AI presents a significant danger to employers “because it makes it easier for a person to give convincing misrepresentations that are difficult to fact-check”.

According to Evans, while technology has assisted employers “to check more forensically the data that’s included in a CV that would otherwise have gone unattended”, development is a double-edged sword, and potential fraudsters are also able to adapt to modern forms of detection.

For “a candidate who, at the very start of the process, has a propensity to want to commit identity fraud, it’s pretty easy to be able to create a driver’s license or a passport or a birth certificate that purports to be someone that they are not”.

Evans noted that WorkPro conducts over 20,000 identify verifications each month as part of background screening.

On detecting anomalies, Evans said a layered and integrated approach is best, with an ideal process including facial verification, a liveness test, police checking, and a government-issued driver’s license or passport.

She said: “Layer that against a document validation service.”

Reflecting that this process “a number of years ago would be seen as very dramatic”, not to mention expensive, Evans asserted that these modern demands mean organisations “are able to adopt and mitigate that risk of candidate fraud within an employment setting”.

However, AI and human verification must work in conjunction.

Fletcher warned that “unguided AI against an ‘AI slop’ is dangerous because it will often pile on and validate untruthful and hallucinated material. It can assist a human review when used carefully, but a human that knows what they’re doing.”

For employers, Evans encouraged the removal of dates of birth, names, and any identifying or qualifying information that may encourage AI bias, “so the right person with the right skills is actually being screened in rather than screened out”.

She said: “Advise the candidates up front in the job application, this is the process that you’re going to be expected to follow.”

Acknowledging the demanding nature of the modern HR role, with technology considerations adding to the strain of complex risk and legislative compliance, reduced headcounts, and demands for fast hiring, lest productivity suffer, Evans asserted there is a place for AI in the hiring process “combined with the appropriate human intervention”, urging employers and HR to use the readily available technology.

“It speeds up the process, but also provides that authenticity and trust and confidence in the process for both parties,” Evans said.

“It’s so easy to adopt technology, and it’s so cost-effective, and it’s so operationally efficient if it’s done in the correct manner.”

Want to see more stories from trusted news sources?
Make HR Leader a preferred news source on Google.