An academic study suggests that AI can extract personality signals from a single photograph and use them to forecast aspects of a person’s career, including salary levels and job mobility. The findings raise major questions about ethics, yet they arrive as employers are increasingly using AI tools in recruitment and promotion decisions.
The research team analyzed nearly 97,000 LinkedIn profile images of MBA graduates from leading U.S. business schools. Using a machine learning model trained to associate facial features with the Big Five personality framework, openness, conscientiousness, extraversion, agreeableness and neuroticism, the system generated personality scores for each individual. Those scores were then compared with detailed career data, including school rank, first-job placement, compensation and subsequent advancement.
The results suggest that the AI-derived personality measures have a statistical relationship to a wide range of labor market outcomes.
But even one of the report’s authors, Marius Guenzel, a finance professor at the Wharton School of the University of Pennsylvania, stress that the study needs to be seen in its proper context. “As a society, we need to be mindful of how this technology is applied, particularly when it relies on traits that individuals can’t easily change,” he said in an interview with a Wharton business journal. “In addition to ethical concerns, there’s also a real risk that meaningful personal growth could be overlooked—and ultimately disincentivized—simply because it doesn’t show up on someone’s face.”
It Pays to Be an Extrovert
Individuals whose inferred traits aligned with well-established occupational demands were more likely to secure higher-paying roles and experience higher salary growth, according to the study. In particular, extraversion and conscientiousness emerged as consistent indicators of upward mobility and compensation increases.
The authors argue that personality has long been recognized as a driver of career divergence among similarly educated professionals, but measuring it at scale has proven difficult. AI-based image analysis, they claim, offers a new way to study those patterns across large populations.
The algorithm underpinning the study was first introduced in 2020 and trained on datasets where participants submitted photos and completed personality surveys. In validation tests, correlations between self-reported personality traits and AI-inferred scores were modest, though they were comparable to other personality measurement techniques used in social science research.
While the researchers describe the technology as offering meaningful predictive power, they did not endorse its use in hiring. Instead, they frame their work as an effort to examine tools that are already gaining traction in industry.
Personality assessments like questionnaires or video analysis are widely used in admissions and corporate screening. Some firms have begun incorporating AI systems that attempt to infer behavioral traits from digital interviews.
Major Ethical Concerns
The ethical concerns raised by the study are, to be sure, major. Because facial features are by their very nature a shallow way to understand someone, critics argue that embedding such analysis into employment decisions risks entrenching bias and stereotype.
The authors say that ignoring the technology will not prevent its adoption. Regulatory frameworks governing AI in hiring remain fragmented, particularly in the U.S. As a result, the researchers argue that scrutiny is necessary to inform policy discussions and public understanding.
In any case, the stakes are significant. Managerial and leadership roles, which rely heavily on soft skills, represent a lucrative segment of the workforce. To fill these jobs, employers often look beyond test scores and academic pedigree to assess personal skills, which some companies might rely on AI to determine.
Skepticism is Required
The idea that a headshot could signal future earnings invites profound skepticism. Previous scholarship has cautioned against overinterpreting machine learning systems that draw inferences from physical appearance. As many have argued, face-based analytics is pseudoscientific, and a study’s technical sophistication does not guarantee real world validity.
The larger question is: how can professionals and companies prevent AI from playing too large of a role in all-important hiring decisions? As AI continues to play a growing role in recruitment and talent evaluation, the issue is no longer whether such tools exist, but how they will be governed.
