Collaborative Research: FW-HTF-RL: Trapeze: Responsible AI-assisted Talent Acquisition for HR Specialists

Human Resources (HR) Specialists fulfill a range of critical staffing functions in organizations. This project focuses on supporting HR Specialists in the technology and “big-box” retail industries, who source and screen candidates for entry- to mid-level positions. These HR Specialists often find themselves under enormous pressure to fill roles, and they turn to automated decision systems (ADS) for managing the meticulous balancing act of talent acquisition: sifting through pools of candidates to find people who meet job requirements and have the “right” culture fit, while adhering to ethical standards and legal compliance. AI models that match and rank candidates are at the heart of these ADS. Poorly designed models can produce incorrect and inconsistent results that fail to match candidates appropriately to job requirements, or that limit the visibility of well-suited candidates. Together, these problems can lead to unaccountable decision-making processes and unfair decision outcomes, harming individual job seekers and members of already disadvantaged communities, and putting employers at risk of litigation.

This project reimagines the role of HR Specialists (future worker), empowering them with the agency to reason about, validate, audit, and influence the ADS-assisted hiring process (future work context). These interventions are supported by a human-in-the-loop framework called Trapeze (future technology) that supports transparent automation in talent acquisition, along with innovative educational materials and methodologies that train HR Specialists to become better informed about AI and accountability in ADS-assisted decisions. Outcomes of Trapeze include open-source software, allowing the broad and diverse community of responsible AI researchers and practitioners to build and evaluate tools for sourcing and screening more effectively. This project also advances the understanding of the behavioral, social, legal, and technical contexts in which HR Specialists in the technology and retail domains make ADS-assisted decisions. Publicly available training materials and methodologies from this project help HR Specialists become more informed, responsible, efficient, and effective in their use of ADS. All shared materials, taken together, serve as a strong blueprint for strengthening accountability in ADS use within other high-stakes sectors of industry.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Award Number: 2614053
Principal Investigator: Frederick Oswald
Funds Obligated: $329,946
State: CA