As AI-powered tools become more prevalent in work and life, consumers and organizations are trying to sort out exactly what AI should be used for. Generative AI can be used to rapidly spin up images and write text, often using a few simple prompts, and promises to help knowledge workers automate critical but time-intensive tasks. However, reducing these sorts of tasks down and running the entirety of their operation through a single AI tool could also create some serious issues.
One discipline that has been identified as a key candidate for AI tools is recruitment. Because traditional recruitment typically involves long, time-consuming searches for candidates using job sites or search engines, the idea of an AI-powered tool that can cut down these searches is attractive to many. On the job candidate side, sending out ten or 20 or 100 job applications is a tedious, repetitive task that offers minimal feedback or support from those you are applying to. An AI tool that can generate cover letters and fill out applications instantaneously would be a huge help to anyone in the midst of a job search.
Nobody can deny the potential usefulness of AI in these and other contexts, but, as with any heavily-marketed tool, many of these supposed benefits conceal some very specific and possibly dangerous downsides. As AI adoption grows, organizations and job seekers will have to sort out just how valuable these tools are, and some may find the risks do not justify the time saved. Here are some of the key risks of using AI in recruitment:
Generative AI risks for recruiters
The key issue with AI for anyone using it is that it is something of a “black box”; in other words, it is a tool that most users only understand based on inputs and outputs. This means, any issues a recruiter might have with it will be hard to detect and even harder to tweak out of the system. This is most evident in the numerous well-documented cases where automated recruiting tools replicated harmful biases based on race, gender, and nationality of applicants. Because these tools draw from extant datasets, if these datasets include a bias for or against someone from a particular group, the AI system will replicate it. Because it is an opaque system to recruiters, they may have a hard time understanding why this is happening, let alone stopping it.
This creates major issues that wouldn’t otherwise be present in the recruitment process. Given that some high-profile examples of AI hallucinations include an AI tool professing to fall in love with users and then spying on them, this represents a major risk for AI use in the highly sensitive recruitment process.
Generative AI risks for job seekers
Limited feedback for calibration
Much like the black box issue described above, one of the primary issues job seekers would face is a lack of feedback. Because AI would likely be used to automate the initial step of job applications, it would be able to apply to a large number quickly. However, given that most applications end with no response, it would be difficult to tell how well the AI is performing. For example, if it sent out 50 applications and only got one response, is that because it sought out jobs that were above the candidate’s level, or because there were issues with the applications or cover letter? The candidate would have to go back through the applications manually to check, which undoes the time-savings advantage AI offers.
Finding a job versus finding the right job
Striking a balance between applying to a large number of jobs to maximize your chances versus applying to jobs you want can be difficult. Because AI leans more towards maximization, a performant AI tool might set up many interviews for you only for you to come away from each unsure if you actually want the job. You cannot easily automate job seeking for roles that meet your ideal preferences or the type of company you want to work for. Even if you are able to calibrate the AI to find these specific roles, it only makes sense to use it if there are so many suitable roles that applying for them all is too time consuming to do manually.
Accounting for generative AI in recruiting
As these tools are tested and incorporated, both recruiters and job seekers will need to think hard about their use and the issues they cause and find novel ways to measure performance. Training with AI usage and prompts is essential, as is manually measuring performance on issues like diversity and application quality. These tools may also have the holistic benefit of reducing the work involved in job applications. As content like cover letters and resumes become grist for automated tools, the people involved can put more emphasis on interviewing, in-person aptitude testing, and other key recruitment tools. In other words, the automation of the recruitment funnel could fundamentally change it, making past practices like writing a cover letter, prompt, or perfectly laid-out resume irrelevant to the actual process.
Get a handle on HR best practices with 501(c) Services
Our team is dedicated to helping nonprofits identify key resources they can use to meet their goals. Because nonprofits often have a difficult time recruiting, our work can help you make key investments in your recruitment and onboarding process and find the ideal candidate for your team. As long-time nonprofit experts, our team is deeply connected with the nonprofit space and finances and can help you find new ways to succeed. If you’d like to learn more, please get in touch.
501(c) Services has more than 40 years of experience helping nonprofits with unemployment outsourcing, reimbursing, and HR services. Two of our most popular programs are the 501(c) Agencies Trust and 501(c) HR Services. We understand the importance of compliance and accuracy, and we are committed to providing our clients with customized plans that fit their needs.
Contact us today to see if your organization could benefit from our services.
Already working with us and need assistance with an HR or unemployment issue? Contact us here.
The information contained in this article is not a substitute for legal advice or counsel and has been pulled from multiple sources.