Algorithms at the Gate: Is AI Reducing or Reinforcing Bias in Hiring?

Artificial Intelligence (AI) and Large Language Models (LLMs) are reshaping recruitment, promising efficiency and objectivity. Tools like one-way video interviews and automated resume screening streamline candidate assessment and aim to reduce human bias. Yet, many job seekers feel that AI-driven tools, rather than leveling the playing field, sometimes reinforce biases and create an impersonal hiring experience (GeekWire, 2024; Washington University, 2024). As employers integrate these technologies into recruitment, understanding how candidates perceive AI’s role in hiring is fairly essential to fostering fair, inclusive, and transparent recruitment practices.

The Rise of AI in Recruitment and Candidate Assessment

AI has quickly become embedded in many aspects of recruitment technology, from screening resumes and conducting video interviews to assessing personality traits through voice and facial analysis. One-way video interviews, in particular, have gained popularity due to their flexibility and scalability. Recruiters benefit from the ability to review candidate responses at their convenience, and companies can efficiently manage large volumes of applications. However, what’s efficient for employers may feel impersonal to job seekers, leaving them feeling like just another data point in a far from perfect system (American Staffing Association, 2023).

Human Perceptions of Bias and Fairness

Research—and even just a quick search on LinkedIn—shows job seekers frequently see AI-driven assessments as uncomfortable and detached. For instance, LinkedIn user Marilyn D. shared her one-way interview experience as “the most uncomfortable of my life,” citing concerns about camera anxiety and potential biases based on appearance (Paulson, 2024). Similarly, Elizabeth Mavor described feeling disconnected, questioning whether a company truly values its people when interactions lack human connection. These stories highlight the tension between AI’s efficiency and the need for empathy in candidate assessment.

Professionals like Richard Chamberlin also point out that traditional interviews allow for richer interpersonal interactions, including reading body language, tone, and pacing—cues that one-way interviews and AI tools often miss. These insights suggest that while AI can optimize certain aspects of recruitment, it may lack the nuance required for a fair and empathetic evaluation of candidates.

Cognitive Biases Embedded in AI

Even AI systems designed to be impartial can inadvertently reflect cognitive biases. Dr. Gleb Tsipursky (2024) identifies common biases, such as confirmation bias, status quo bias, and loss aversion, which can impact AI-driven hiring decisions. For example, confirmation bias might lead algorithms to favor candidates with profiles that mirror historically successful hires, reinforcing existing hiring patterns rather than promoting diversity. Status quo bias may drive companies to adopt AI without examining its potential for inclusivity, especially as these tools become embedded in recruitment workflows. Such biases within AI systems, combined with candidates’ skepticism, create challenges for organizations aiming to offer fair and balanced hiring experiences (Tsipursky, 2024).

Research on AI Bias in Resume Screening

Recent studies indicate that AI in recruitment can, in some cases, perpetuate biases. For instance, Washington University’s research suggests that AI-based resume screening systems often favor white and male candidates, raising concerns about inclusivity (Washington University, 2024). GeekWire’s analysis aligns with these findings, showing that AI may overlook qualified candidates from underrepresented backgrounds, potentially due to biased training data (GeekWire, 2024). For job seekers, these findings reinforce a perception of unfairness, undermining trust in AI-driven recruitment processes.

Addressing Fairness in AI-Driven Recruitment To bridge the gap between AI’s potential for objectivity and the reality of candidates’ experiences, companies need to address fairness and transparency head-on. Here are some practical steps:

• Implement Feedback Mechanisms: Create opportunities for candidates to provide feedback on their AI-driven experiences. This input can be invaluable for refining tools and demonstrating to candidates that their experiences are valued.

• Use Diverse and Representative Training Data: Training AI models on diverse data sets is essential for reducing bias and providing fairer assessments (American Staffing Association, 2023).

• Maintain Human Oversight: AI is a powerful tool, but human judgment should remain central in hiring. Incorporating human review at key stages of the process can catch potential biases and create a more balanced assessment.

• Transparency and Communication: Clearly communicate how AI tools are used in recruitment and how decisions are made. Transparency fosters trust and helps candidates feel respected and valued.

• Invest in Ethical AI: Developing guidelines for ethical AI use in recruitment, including bias mitigation and data privacy, is an important step. While these guidelines are likely to be developed by IS/IT and legal teams, HR and Talent Acquisition (TA) should still play a central role in advisement to keep people-focused values at the forefront during development. Regular audits should help ensure these practices evolve alongside advancements in technology.

Conclusion: Building a Fairer Future in AI Recruitment

As AI and LLMs continue to shape recruitment, prioritizing fairness, transparency, and empathy is a non-negotiable for employers to build trust in the hiring process. Organizations have a unique opportunity to reshape the candidate experience by addressing implicit and explicit biases within AI-driven processes, ensuring these tools serve all job seekers equitably. While AI offers undeniable efficiencies, balancing these tools with human oversight can strengthen trust, create a more inclusive hiring experience, and set the stage for a positive employment relationship from the very beginning.

Call to Action: Your Help Is Needed

Your insights can help make a difference. If you have are a job seeker, or have looked for a job in the last 12 months, please consider contributing to a study on LLM's and AI's effects on job search outcomes. I need 43 more participants to complete important research as part of my dissertation in I/O Psychology on how AI influences job seekers’ experiences and job search outcomes. Your participation could help drive meaningful change in recruitment practices. Please take the survey here.

Thank you.

References

• American Staffing Association. (2023). AI recruiting tools. Retrieved from https://americanstaffing.net/posts/2023/09/07/ai-recruiting-tools/
• GeekWire. (2024). AI overwhelmingly prefers white and male job candidates in new test of resume screening bias. Retrieved from https://www.geekwire.com/2024/ai-overwhelmingly-prefers-white-and-male-job-candidates-in-new-test-of-resume-screening-bias/
• Paulson, C. (2024). LinkedIn post on AI and one-way video interviews. LinkedIn. Retrieved from https://www.dhirubhai.net/posts/colleenmpaulson_lots-of-us-are-figuring-out-how-we-can-leverage-activity-7259550030728937473-nSpr?utm_source=share&utm_medium=member_desktop
• Tsipursky, G. (2024). Leading the generative AI transition beyond cognitive biases. Vistage Worldwide, Inc. Retrieved from https://www.vistage.com/research-center/business-operations/business-technology/20241011-generative-ai-transition/
• Washington University. (2024). AI bias in resume screening: race and gender impacts. Retrieved from https://www.washington.edu/news/2024/10/31/ai-bias-resume-screening-race-gender/. Vistage Worldwide, Inc.
• Washington University. (2024). AI bias in resume screening: race and gender impacts. Retrieved from https://www.washington.edu/news/2024/10/31/ai-bias-resume-screening-race-gender/