The hidden costs of AI-driven hiring decisions

An insightful exploration into how today's recruitment technologies may be shaping the future of work, for better or worse.

A recent survey by IBM, encompassing over 8,500 IT experts worldwide, revealed that nearly half (42%) of organizations are integrating AI-based screening tools within their recruitment and HR practices to enhance efficiency. Despite the widely acknowledged issue of bias in AI and the common belief that it may perpetuate human prejudices, it's peculiar that many leaders still believe that AI-driven recruitment could address and rectify bias within the hiring process. But how did we get here?

Facing the algorithm: the realities of AI-driven recruitment

The average cost of a bad hire is around $15,000. Traditionally, mirror image recruiting was blamed for introducing bias into the hiring process. As a result, large enterprises turned to technology to solve the problem.

A few years ago, Unilever began employing AI to analyze candidates' facial and linguistic responses to modernize recruitment by leveraging AI and facial recognition technologies. However, the allure of technology as a bias-free selector is misleading. The technology developed by HireVue, promising to eliminate human bias, underscores a stressful interview process where machines scrutinize every facial twitch and verbal nuance.

This digital approach, while innovative, doesn't guarantee fairness; it's fundamentally rooted in mathematics, susceptible to the "garbage in, garbage out" principle.

Algorithms depend on the data they're fed, and without a diverse dataset, they risk perpetuating the biases they aim to eliminate. Ignoring this can lead to the exclusion of talented individuals who fall outside conventional norms, ironically due to the human prejudices embedded within AI training data.

AI in recruitment: understanding automated video interviews

Automated Video Interviews (AVIs), also known as on-demand or asynchronous interviews, represent a leap into the future of recruitment, inviting selected candidates to engage with platforms like SparkHire, myInterview, and HireVue. These platforms prompt applicants to record and submit video responses to pre-set questions, with AI algorithms analyzing facial expressions, keyword usage, and tone of voice.

The digital approach promised efficiency and objectivity by processing more applicants with fewer resources and potentially reducing implicit bias. Candidates wanting to beat the algorithm are encouraged to practice speaking into a camera, which can help simulate the interview environment, making it less alien during the actual process. Additionally, ensuring a professional setting – adequately lit and free from distractions – alongside maintaining direct camera eye contact can significantly improve one's presence on screen.

As AVIs become more prevalent, understanding their mechanics and mastering the art of digital presentation will be indispensable skills for job seekers navigating the evolving recruitment landscape. But will it merely champion candidates who can talk BS confidently about anything like ChatGPT?

Pseudoscience in recruitment: unpacking AI's dubious foundations

Anthea Mairoudhiou, previously employed as a makeup artist at MAC, found herself jobless in 2020 after a video interview. Despite achieving a perfect score on her performance, Mairoudhiou was dismissed due to the AI-driven HireVue system penalizing her for her body language during the assessment, illustrating the technology's potential for misinterpretation and the profound consequences on individuals' careers.

Unfortunately, this was not an isolated incident. When Reverend Paul White tweeted that his daughter's interview for a graduate job would be with an AI program, it went viral, with 2 million people sharing their support.

Hilke Schellmann, an Emmy award-winning investigative reporter, recently told the Guardian that many AI hiring tools are grounded in dubious pseudoscience. For example, the belief that voice intonation can determine your success in a job not only lacks credibility but also possesses the potential to discriminate.

Relying on algorithms and keyword matching can lead to a narrow focus, overlooking applicants' nuanced capabilities and potential. This mechanical approach raises questions about the actual efficiency of such systems, suggesting that what is gained in speed may be lost in depth.

Personal touch vs. automation

The importance of personal connections and networking in the job application process cannot be overstated. Direct communication with hiring managers or department heads often leads to better outcomes for candidates, underscoring the limitations of automation in understanding and valuing human potential.

Sir Richard Branson's reflection on the liberation his dyslexia afforded him in the real world, transforming perceived obstacles into innovative solutions, underscores a broader narrative in the tech and business world. Other leaders like Peter Thiel have highlighted the unique advantages conditions like Asperger's offer in industries that thrive on forward-thinking and innovation.

The recognition of neurodiversity as a catalyst for creativity and problem-solving challenges the narrative around AI in the hiring process. Using facial recognition to assess candidates' language, tone, and facial expressions risks sidelining uniquely talented individuals by adhering to a narrow, flawed criterion.

The changing landscape of job applications

The shift towards automated systems has transformed job applications into impersonal transactions, stripping away the human engagement that can be crucial in identifying the right candidate for a role. This evolution raises concerns about the recruitment industry's direction and job seekers' experiences navigating these automated gateways.

Candidates have resorted to strategies to "game" the system, from keyword optimization to direct outreach to bypass AI filters. These tactics highlight a broader issue within the recruitment process, where the ability to navigate automated systems becomes as important as the qualifications and skills of the applicants themselves.

Critics argue that the recruitment industry's reliance on AI tools, driven by efficiency and cost-cutting motives, comes at the expense of truly understanding and valuing candidate potential. This critique calls for reevaluating priorities within the industry, questioning whether the pursuit of automation undermines the goal of finding the best fit for a position.

The debate over AI-driven hiring touches on philosophical questions about the nature of the "best" candidates and whether algorithms can objectively determine such a concept. This reflection challenges the recruitment industry to consider broader definitions of suitability and potential beyond what AI can quantify and analyze.

Call for a balanced approach

Businesses are stripping the human element from HR processes by employing AI to craft job advertisements, sift through resumes, evaluate candidates, and analyze body language and facial expressions during video interviews. Ironically, this cold process reveals more flaws and warning signs about the organization than the candidate.

The key to unlocking genuine innovation and addressing bias and diversity should not be leveraging technology to find a homogeneous ideal candidate. If only we could use tech to discover and celebrate every candidate's unique kaleidoscope of human talents and abilities. Now, that would be progress.