Spotting AI-Generated Resumes       

(And Does it Matter?)

It’s not news that AI plays a role in the evaluation of resumes, but it hasn’t always been successful in the past. In 2018, Amazon had to scrap an AI-powered resume screening tool when its bias was revealed: it downgraded women applicants since it was trained on data from an era where men dominated tech fields. Another study from 2022 discovered that AI resume scoring could often be affected by irrelevant factors; for instance, identical resumes would get different scores based purely on their choice of file format. Since AI “learns” from patterns in past data, it makes sense that it might replicate previous biases – but that’s not ideal for accurate and thorough resume evaluation.

Today, more conversations are happening around the ethical and practical concerns of AI, especially now that there’s a higher risk of candidates using it to generate their resumes. If your team is worried about bot-generated resumes, you’re probably focused on two concerns: how to spot them and how to react to them.

How to Identify AI Resumes

When faced with a resume you suspect was AI-created, your first instinct may be to input a resume into a text bot, like ChatGPT, and ask if the document is AI-generated, but that won’t actually give you useful results. Just ask one professor who made national news when he reportedly did just that with his students’ essays and falsely accused them of plagiarism! Remember: AI text bots don’t actually “know” anything; they simply sift through enormous quantities of data and predict the most likely answer to a given query. In other words, it’s not evaluating the text you input; it’s just searching for the most common answer to “Is this AI-generated?”!

The human eye can spot certain hints that a resume is AI-generated, such as generic but formal language, non-specific or vague details, a lot of repetition, and a lack of “voice” or personality. While it’s not a 100% guarantee, some of these factors can raise the possibility of an AI-generated document. So, if you do have a suspicious resume in front of you, what should you do?

How Much Does AI Matter to You?

We now live in a world where AI technology exists and is easy to use, so we have to figure out how much weight to assign. A resume is, after all, just one part of an application, so the question becomes: is the use of AI in its creation a “yellow flag,” a total dealbreaker, or irrelevant? It all depends on the job and organization.

Integrity matters in any job, of course, but some roles require more sensitivity and ethical guidance than others. For example, suppose a role involves handling sensitive data, producing or evaluating unique written work, or significant creativity or strategic planning. In that case, it’s more of a warning sign to see an AI-generated resume (as opposed to a role that focuses more on pure numbers or data analysis). Likewise, an organization bound by nuanced and detailed ethical guidelines may not be as likely to hire someone with a suspected AI-generated resume.

The level of the role matters, too. Screening every entry-level hire for AI-generated resumes may not be worth the effort. On the other hand, when hiring for higher stakes and higher-level positions, your team may see AI usage as a bigger red flag. In these roles, it’s more important to know that the successful candidate truly has all the skills necessary – including writing and communications skills – to succeed.

Going Beyond the Resume

Resumes should always be just one part of the candidate evaluation process. Today, these non-resume aspects are even more relevant, as they can paint a fuller picture of a candidate without potential AI interference. If a seemingly strong candidate has a suspected AI resume but hasn’t raised any deal-breaking issues, consider these other factors:

  • Read cover letters closely. Because a cover letter should be more personal (as well as a sample of writing abilities), an AI cover letter is more alarming than an AI-generated resume. Look for AI “tells” in the cover letter, and exercise more caution if it seems suspicious, too.
  • Use practical applications where possible. Hiring for some types of roles may include looking at past work samples or giving candidates practical exercises to complete. The results of these can provide valid insights into applicable skills without the risk of AI – just be sure that any “assignments” submitted don’t feature AI tells, too.
  • Ask more behavioral and scenario questions. Behavioral questions allow interviewers to understand, in a candidate’s own words, how they would respond or have responded to certain situations. In turn, this can give much stronger, more nuanced insights into their thought processes, separating well-rounded candidates from those with interesting resumes – AI-generated or not.

There’s no putting the AI genie back in the bottle, and hiring teams and recruiters will have to find new ways to identify the true standout talent from a potential sea of AI “assisted” applications. Fortunately for all of us, we already have the tools and techniques to do this. We just have to refocus our efforts on evaluating the factors that can’t be AI-generated and spend more energy considering the details and understanding people as individuals – that way, we can truly separate exceptional, promising talent from those relying on an AI boost.