How candidates are using AI to flood the job market with resumes

When speaking with other HR and recruiting professionals, they are generally more concerned about AI than they are excited. Although we all use AI to varying extents, it hasn’t been super transformative in our work, representing a time and effort-saving tool. Instead, we are more concerned about how candidates are using AI in their job search, largely to our detriment.

Instead of carefully crafting a personalized resume or diligently filling out application forms, ChatGPT has largely replaced this effort. Even if AI were used to completely generate the resume, candidates almost certainly will use tools like Grammarly to curate their resumes.

But how exactly are candidates using AI tools in their job search, and what are the negative effects of this practice for HR practitioners? We have spent significant time researching this and looking at how AI is being used in recruitment.

Mass Produced Written Content

Historically, organizations have used resume sifting, cover letters, and application forms as a soft measure of written ability, helping to identify candidates with strong verbal reasoning. Additionally, targeted personal statements and thoughtful cover letters implied a strong motivation to work at the organization, as candidates would be required to invest significant time and effort into applications. Although by no means a perfect indicator of performance, this approach at least offered some rationale for shortlisting, and has been widely used by companies the world over.

However, ChatGPT now enables candidates to mass-produce CVs and cover letters (as do other Gen AI tools like Kickresume, Rezi, and Enhancv). Gen AI can also write well-structured application form responses and even apply to jobs for you (using tools like LazyApply, Sonara, or Massive). So CVs and application forms are no longer a viable selection criteria. Candidates can simply mass-produce this content with ease, avoiding any spelling or grammatical mistakes in the process. This is particularly frustrating for organizations who formally grade applicants based on their application form responses, which is common in many graduate-level positions that receive high volumes of applicants.

More pernicious, however, is the fact that candidates who use AI are able to apply for a huge number of jobs, far more than would otherwise be possible. Although perhaps a minority of candidates are using AI in this way, a large portion of genuine applications you receive will likely use AI-written content. These applicants are less invested in the process, and may not even turn up for interview.

Recommendations for HR

First and foremost, you need to abandon the idea that old-school resume sifting works. Instead, I strongly recommend using a simple application form with qualifying questions, i.e., do you have a relevant qualification, do you have relevant work experience, do you have the right to work in the country, etc. If they answer each question appropriately, then just progress them to the next stage in the recruitment process without looking at their resume or any cover letter. There is almost nothing to be gained from reading ChatGPT output masquerading as a cover letter, so save everyone the effort and don’t ask for one. Although this can be a difficult transition for organizations who are used to working closely with resumes, as long as you can get the information needed from a simple form, there is nothing else to be gained from reading their resume, so don’t waste your time.

Instead, I recommend finding a better way to shortlist when required. Particularly when it comes to early careers, I recommend using pre-employment tests such as ability tests or personality questionnaires. These are harder for AI to cheat, and there are even some which guard against AI cheating, such as those provided by Test Partnership. Such assessments are an effective filter for high volumes of applicants, allowing you to create a quality shortlist of high-potential applicants. Additionally, I would recommend doing two-way video interviews, as opposed to one-way (also known as asynchronous) video interviews. This is because candidates on asynchronous video interviews can easily generate a script using ChatGPT and just read it aloud.

Conclusions

Although it can be frustrating to change your recruitment processes reactively, the overall goal of recruitment remains the same, i.e., hiring the best people for the job. HR professionals should take this moment to pause, reflect, and reevaluate what they are trying to accomplish in their recruitment processes, and adapt accordingly to changes in the job market. If written content is losing relevance as a screening tool, and the total number of applications is on the rise, you need an alternative approach to shortlisting, and plenty of options are available (notably ability tests or personality questionnaires). The sooner you make the transition, the sooner you can refocus your time on the ultimate goal of recruitment, helping you to maximize the quality of hire while streamlining the process.