"Hiring Black Box" – Are You Better at Job Recruitment Than an AI? 

Summary:

The “Hiring Black Box” installation at the EASST-4S 2024 conference explored the role of AI in recruitment and its potential biases. This is the flagship conference in Science and Technology Studies, with almost 4000 attendees. Participants could engage in an interactive “making and doing” session where they compared their human candidate selections with those recommended by AI, highlighting discrepancies and fostering discussions on fairness. The making and doing session emphasized understanding AI decision-making, recognizing human biases, and promoting inclusivity in hiring processes. Feedback from participants provided valuable insights for improving AI-enhanced recruitment systems and recommenders. Overall, the making and doing session underscored the need for transparent, fair, and inclusive AI in human resource management and was a creative way to explore these issues. This blog post recaps the making and doing session led by Shan Wang, Silvia Ecclesia, Roger Andre Søraa & Mark Kharas from NTNU, with technical support by BIAS-project partners from Bern University of Applied Sciences and Digiotouch, with overall design by LOBA. The installation was also a finalist for the Making and Doing 2024 award.

Introduction

The rapid advancement of Artificial Intelligence (AI) has revolutionized numerous sectors, including human resource management. As AI becomes increasingly integrated into the hiring process, questions regarding the fairness and bias of these systems have come to the forefront. The “Hiring Black Box” making and doing session, part of the 2024 EASST-4S conference, aimed to explore these critical issues.  It examined the impact on understanding AI’s role in recruitment. Artificial Intelligence is increasingly utilized in human resource management for selecting, training, and evaluating workers. This making and doing session explored the fairness of the recruitment process and how AI can either mitigate human bias or exacerbate discrimination.

Installation and Session Structure

 The “Hiring Black Box” making and doing session was a physical installation with a technical setup. Participants experienced an initial screening process in hiring activities from both the AI’s perspective and their own. Initially, they read a job description and indicated the ranking criteria they deemed important, such as previous experience, education, and language skills. The AI then made an assessment based on the human input, while the participants made their own selections and rankings of the job candidates. After exiting the black box, participants compared their choices with those of the AI to see if and how they differed. Finally, participants provided feedback on the activity to help the BIAS project rethink AI-enhanced hiring processes. Step by step, it worked as follows:

1. Introduction and Explanation: The making and doing session began with an introduction to the concept and objectives of the “Hiring Black Box”. Participants were briefed on the increasing the role of AI in recruitment and the potential biases that could arise.
2. Job Description and Criteria Selection: Participants were given a job description and asked to select ranking criteria they considered important for the job. This could include previous experience, education, language skills, and other relevant factors.
3. AI Assessment: The AI system made an assessment based on the input provided by the participants. This step illustrated how AI algorithms process and evaluate candidate information.
4. Human Selection: Participants then entered the “black box” to make their own selection and ranking of the job candidates. This allowed them to experience the recruitment process firsthand and understand the complexities involved.
5. Comparison and Discussion: After exiting the black box, participants compared their choices with the AI’s selections. This comparison highlighted any discrepancies and sparked discussions on the reasons behind these differences.
6. Feedback and Reflection: The final part of the making and doing session involved gathering feedback from the participants. They reflected on the activity and provided insights on how the BIAS project could rethink AI-enhanced hiring processes.

Impact and Insights

The “Hiring Black Box” making and doing installation provided participants with a deeper understanding of AI’s role in recruitment and the potential biases that can arise. Key takeaways from the making and doing session were:

1. Understanding better AI’s Decision-Making: Participants gained insights into how AI algorithms make decisions based on the criteria provided. This understanding is crucial in assessing the fairness and transparency of AI systems.
2. Highlighting Human Bias: By comparing their selections with the AI’s choices, participants could identify their biases and consider how these might influence hiring decisions. This self-awareness is essential for reducing bias in the recruitment process.
3. Encouraging Critical Thinking: The making and doing session encouraged participants to critically evaluate the use of AI in recruitment. It highlighted the need for continuous monitoring and evaluation of AI systems to ensure they are fair and unbiased.
4. Promoting Inclusivity: Discussions during the making and doing session emphasized the importance of designing AI systems that promote inclusivity and diversity. Participants recognized the need for diverse data sets and inclusive algorithms to prevent discrimination.
5. Feedback for Improvement: The feedback collected from participants provided valuable insights for the BIAS project. These suggestions will help refine AI-enhanced hiring processes and make them more equitable.

Conclusion & final thoughts

The “Hiring Black Box” making and doing session was a thought-provoking and impactful session that explored the intersection of AI and human bias in recruitment. By engaging participants in an interactive and reflective process, the making and doing session shed light on the complexities of AI-enhanced hiring and the importance of fairness and inclusivity. As AI continues to evolve, it is essential to address these challenges and ensure that technology serves as a tool for equitable and unbiased decision-making. The insights gained from this making and doing session will undoubtedly contribute to the ongoing efforts of the BIAS project in rethinking AI-enhanced hiring processes.

The EASST-4S 2024 conference provided a unique platform for this making and doing session, fostering a collaborative environment where participants could share their experiences and perspectives. As we move forward, it is crucial to continue these discussions and work towards developing AI systems that are transparent, fair, and inclusive. The “Hiring Black Box” making and doing session was a step in the right direction, and its impact will resonate with participants as they navigate the evolving landscape of AI and human resource management.