Mustafa Akben, a faculty member at the Martha and Spencer Love School of Business, has created a machine learning model that is effective in assessing job performance and leadership potential simply by analyzing emails.
share:
Is it possible to use artificial intelligence to identify job candidates with leadership skills simply by reading email replies to hypothetical situations at work?
This was a challenge for Prof. Mustafa Akben. Over the winter months, he developed an AI model that was more accurate than his 20 other entries in an international competition hosted by the Society of Occupational and Organizational Psychology.
The SIOP 2023 machine learning competition focused on natural language processing. This is a subfield of artificial intelligence focused on enabling computers to understand, interpret and generate human language in recent headlines through its use in ChatGPT.
Here’s how the contest works:
At the Job Evaluation Center, job seekers were asked to respond via email to hypothetical scenarios received during a full-day simulation. The way they analyzed the scenarios, and the tone and content of their responses, were scored by a panel of human judges.
Machine learning competition teams were given a spreadsheet with prompts and job seeker responses. However, it only received a limited number of scores. Task: Determine how a human judge scored her second batch of responses by building a model that looks at the first batch of scores.
Akben studies the impact of artificial intelligence on human productivity and creativity. The AI model he developed performed best overall in predicting how judges rated the second batch of candidates at the public panel, held in Boston in April. provided virtual presentations to in-person and online attendees at his 2023 SIOP Annual Conference.
“I am very excited about this win, both personally and professionally,” Akben said. “It is amazing to see all the hard work, dedication and persistence I put into researching AI, machine learning and organizational behavior paying off. Reinforcing my dedication to excellence, this is a great example of the value of lifelong learning and the pursuit of success in your field of interest.
Moreover, it demonstrates how interdisciplinary research drives innovation and tackles some of the most complex challenges facing the modern workplace today. “
This doesn’t mean AI is about to overtake human assessments of job performance or leadership traits, Akben said. His model did an excellent job of predicting, on average, how candidates would perform in the simulation, but the resulting effect sizes were only moderate.
Akben said he hoped the effect sizes would be larger, and that part of the continued improvement in artificial intelligence is also in eliminating bias. Citing the parameters of competition, he noted that machines may not be able to explain that English is someone’s second language. It may count syntax or grammatical errors for those who
Akben joined Elon University’s School of Business and Entrepreneurship in 2022 after earning a BA from Marmara College, an MS from the University of Pennsylvania, and a PhD from Temple University.
His academic interests in positivity, creativity, social network analysis, biohacking, and machine learning are reflected in the many experiential learning assignments he has already created for his management course principles at university.
Akben said: I would love to collaborate on his AI-related projects, including workplace and educational applications, creativity, and positivity with students and faculty at Elon University. “
