HireVue's AI Explainability Statement

In 2022, HireVue proudly released our AI Explainability Statement—the first of its kind in the HR industry. This statement, reviewed by the UK’s Information Commissioner’s Office (ICO), set a new standard for transparency in AI-powered hiring technologies, offering a clear, accessible explanation of how AI works within the HireVue platform. 

This bold move demonstrated our unwavering commitment to ethical AI use, helping customers and the public understand the logic behind our AI-driven assessments. Publishing this much proprietary information was a significant departure from the traditional industry standard of fiercely protecting intellectual property, but the payoff has been pivotal in building trust amongst our stakeholders. 

Fast forward to today, and we recently expanded on the project by adding a second version of the statement - one that is updated to cover the next generation of our hiring platform (T2O), as well as overall advancements in how we are thinking about technology. Explainability work is not just a one-time declaration but an ongoing dialogue that keeps pace with the future of AI in hiring. The recent updates give our customers, regulators, and candidates an updated resource for reference.

Both documents, the 2022 Explainability Statement and the 2024 Explainability Statement are publicly available. Below is a summary of the major updates of the 2024 T2O version of our explainability document:

1. Platform Updates: The 2024 document introduces the Talent to Opportunity Platform™, highlighting enhancements in functionality, such as hosting over 40 million video interviews and 200 million assessments globally. 
2. AI Techniques and Components: The 2024 version discusses the development and fine-tuning of our updated AI competency models which are used to evaluate candidate responses to interview questions. These models expand the diversity and global representativeness of our AI models. The 2024 version also includes additional details about our Game Based Assessments (GBAs). 
3. Ethics and Consent: The 2024 statement elaborates further on the process, including how candidates may opt out of AI evaluations and still participate in the hiring process. 
4. Testing and Validation: The 2024 version contains a more detailed explanation of how our AI is tested, including the use of multipenalty optimized models and ongoing bias mitigation strategies. 
5. Stakeholder and Accessibility Focus: The new document includes a more detailed discussion on stakeholders (customers, candidates, and external groups) and accessibility concerns, such as accommodations for candidates with disabilities.

By continually refining our technology, embracing the latest research, and prioritizing ethics, we are driving the future of fair hiring technology on our mission to connect talent to opportunity. 

About the Authors

Dr. Nikki Dudley, VP of Assessments R&D at HireVue

Nikki Dudley, Ph.D. is the Vice President of Assessments Research and Development at HireVue where she leads a team responsible for researching, prototyping, and bringing new products to market that incorporate the latest advances in selection science.  She holds a Ph.D. in Industrial Organizational Psychology from George Mason University.

Dr. Lindsey Zuloaga, Chief Data Scientist

Lindsey Zuloaga, Ph.D., is the Chief Data Scientist at HireVue, leading a team that develops machine learning algorithms to predict job-related outcomes. She holds a Ph.D. in applied physics from Rice University. In addition to numerous academic publications, she is a prolific public speaker and has contributed to publications such as Forbes and The Wall Street Journal. 

Dr. Josh Liff, Director, Product Science

Josh Liff, Ph.D., is the Director of Research and Development, Product Science at HireVue, where he integrates machine learning and IO Psychology to create advanced assessment solutions and talent technology products. He holds a Ph.D. from Colorado State University in Industrial/Organizational Psychology and has published in the Journal of Applied Psychology.