With almost 50 PhD-level psychologists on staff, Modern Hire works diligently to develop beneficial AI and deploy it in a safe, ethical, and humane manner for organizations and for individuals.

To that end, we’ve developed six core principles for Ethical AI and Algorithms that we rigorously follow in developing our software and assessments. We believe these principles establish a sound footing for industry practice, based on research and decades of experience in measuring human performance.

 

Modern Hire’s Six Principles for Ethical AI and Algorithms in Hiring

1. AI should benefit both organizations and individuals, including candidates.


2. AI should avoid and verifiably reduce bias.


3. AI must be as transparent as possible in its operation.


4. AI must respect the privacy of candidates and organizations.


5. AI is accountable to candidates and organizations, meaning its outputs must make sense and be auditable.


6. AI research and applications must uphold scientific standards, including the use of representative samples and peer review.

Modern Hire’s Principles for Ethical AI and Algorithms lead to several practical applications across our product set. Among these:

• We collect only job-relevant responses from candidates that are specifically intended to be used in the hiring decision and are under the control of the candidates. Modern Hire does not analyze data that is not intended by candidates to be used in the hiring process, including social media data, facial recognition data, and how a person sounds when talking (i.e., prosody).


• We conduct rigorous criterion-related validity studies on new assessment content. Modern Hire’s assessments are designed and carefully developed to be highly job-relevant and are validated to ensure they predict critical job and organizational outcomes while minimizing adverse impact against protected classes. Full technical reports are available describing the process and validation results.


• We inform candidates about our use of algorithms and encourage transparency industry-wide. Modern Hire welcomes regulation that requires disclosure to candidates about the use of AI, and we encourage informed consent throughout the process to gain candidate trust. Our scientists have conducted peer-reviewed research on how to best present technical information to candidates to maximize their understanding when AI is being used.


• We regularly monitor our tools for anti-bias. As candidate demographics and data evolve, bias can develop in tools previously found to be bias free. Modern Hire works with clients regularly and systematically to analyze hiring data and to make adjustments that ensure bias is minimized.  Additionally, Modern Hire has developed a patent-pending approach to multi-outcome optimization that allows validity and fairness to be optimally balanced in a set of data. This ground-breaking technique will help ensure that both effectiveness and bias mitigation are maximized in selection decisions.


• We promote glass box model transparency over black box approaches. Rather than simply allow algorithms to operate in an opaque or black box manner, Modern Hire works to ensure clarity in model design. It is vital that the inner workings of complex algorithms be understood as making rational sense and that they do not base scores on spurious relationships or potentially biasing factors. Modern Hire’s psychologists work to ensure that the factors driving hiring predictions are logical and explainable to outsiders.