Generative AI tools are obviously and profoundly changing the nature of hiring. Applicants are being asked to demonstrate their skills in new ways, such as through skills-based resumes and interactive employment tests, games, and chatbots. Are the skills that are supposed to be measured actually being measured by GenAI? That is material for a future blog (or two).
But for now, let me point out three trends in GenAI hiring that, by any reasonable standard, are frustrating to both applicants and organizations:
Applicants can mass-apply to jobs with the click of a button. As a result, organizations are being flooded with job applications (e.g., see here and here).
A vicious GenAI application cycle results: Organizations are flooded with applications → meaning that each applicant has to apply to more organizations to land a job → which only increases the flood of applications.
Increasingly, job applicants are feeding job postings into a GenAI (LLM) tool to draft their cover letters, polish their resumes, and complete other application materials. As a result, job applicants/applications look increasingly similar (harder to distinguish).
Put these three trends together — and Houston, we have a problem! How can this problem be solved? If you ask the tech world, the answer is to use more GenAI products!? For example:
Companies sell services using GenAI to triage a massive pile of application materials (think ATS-embedded rankings, talent matching, conversational screeners). From that, a subset of applicants will be recommended for further consideration. But consider the following thought experiment (or actual experiment, perhaps!): What if an organization followed up on a randomly selected subset of applicants of the same size instead? Would the applicants ultimately hired be any worse (and by what metrics would you decide that)?
“Humanizer” GenAI tools attempt to make applicant materials look less like an LLM (attempting to escape detection by hiring organizations who…buy more AI services to try and detect LLM-generated applications).
This AI arms race continues — but perhaps principles will prevail over time. And by principles, I mean the SIOP Principles for the Validation and Use of Personnel Selection Procedures (2018, free PDF download) that provides important guidance for upholding the reliability, job-relevance, validity, and fairness of employment tests and personnel selection systems — whether GenAI is involved or not.