Harqen.AI, a leading provider of AI-powered evaluation technologies for both corporate recruiting and college admission processes, today announced preliminary findings about the accuracy of the company’s AI technology by acclaimed Purdue University researcher and Industrial and Organizational (I/O) Psychologist Michael A. Campion, Ph.D.
Dr. Campion, who is among the 10 most published and cited authors in the fields of I/O Psychology and Human Resources management, was hired by Harqen as a third-party consultant to help guide its ethical AI development and provide audits of the company’s AI models and practices.
“Harqen’s results have shown that its AI technology is as accurate or better as the humans who currently make selection decisions such as recruiters, admissions advisors, interviewers, managers, or other hiring officials,” said Dr. Campion.
Dr. Campion continued: “Accuracy is measured as both the percentage of agreement in actual accept/reject decisions made by the humans or the scores the humans assign to the candidates (if the organization uses a scoring system). In terms of the accuracy of agreements, for example, if human hiring officials agree on candidate quality 75% of the time, Harqen’s AI will agree with the human judgments 75% of the time or better. If human hiring officials agree 90% of the time, then Harqen’s AI will agree with them 90% of the time or greater. In terms of the accuracy of scores, for example, if the scores of human hiring officials are highly correlated with each other, then Harqen’s AI scores will be highly correlated with the human scores.”
Harqen uses the latest artificial intelligence tools, including advanced text analytics, decision trees, and other machine learning algorithms to score candidate narrative answers to interview questions, descriptions of work experience, type of education, degrees, grades, scores on tests, and other background information available in resumes, job applications, or college applications.
For its AI-powered interviewing solution, it is important to note that Harqen transcribes all interview answers into text before any analysis and does not utilize controversial assessment methods such as facial analysis, voice inflection analysis, self-assessed personality tests, or gamification.
“For the hiring process, Harqen offers a highly accurate solution to automate applicant screening. The results are vastly improved efficiencies that allow employers to identify top talent quickly while also giving candidates a better and faster recruiting experience,” said Harqen Chairman and President, Stuart Olsten.
Mr. Olsten added: “In the area of college admissions, particularly with graduate programs, we are providing quick and timely results that support academic administrators in making their admissions decisions. In addition, we are in early testing of new analysis and models with our academic partners that will better predict student outcome success, such as program completion and grades.”
Harqen’s AI can effectively model (statistically replicate) the human decision-making process. All that is needed is to provide the Harqen AI tools with the same information as is used by the humans (e.g., applications, resumes, interview answers, etc.) and the decisions or scores of the human hiring officials for a reasonably large sample of past candidates. A Harqen AI model can then be developed and used in the future to select candidates without the human effort normally required to do so at much lower cost and much faster, and at the same or better accuracy.