Advances in artificial intelligence are now intersecting with the labour market in a way that could fundamentally alter hiring and career progression processes. Researchers from institutions such as Wharton and Yale have developed a novel methodology that extracts the Big Five personality traits – openness, conscientiousness, extraversion, agreeableness, and neuroticism – from facial images using AI. This approach, dubbed the Photo Big 5, uses LinkedIn data and has already demonstrated its ability to predict critical labour market outcomes, including salary, seniority, and job mobility.
The study, which analysed over 96,000 MBA graduates from top-tier business schools, found that personality traits derived from facial images can predict outcomes with comparable accuracy to traditional measures like GPA or test scores. For instance, conscientiousness, a key trait for academic and workplace success, was strongly linked to higher rankings of MBA programs attended. Extraversion, on the other hand, emerged as a critical predictor of initial salary levels. The implications are profound: these traits, which can be detected at scale without subjective surveys, offer employers a powerful new tool for assessing potential candidates.
However, the study also highlights significant ethical concerns. By relying on immutable facial features, this technology introduces a new form of statistical discrimination. While traditional metrics like academic achievements can, at least in theory, be improved through effort and learning, facially inferred traits could bypass these aspects, potentially reinforcing biases. In addition, these methodologies risk commoditising human individuality, reducing individuals to a set of characteristics that might not capture the full complexity of their personality or potential.
Another challenge lies in the inherent variability of these predictions. While the Photo Big 5 correlates well with economic outcomes, it does not account for the full range of human behaviour. Much like traditional metrics, its predictive power is limited, leaving substantial variability unexplained. This raises questions about its suitability as a standalone tool in recruitment or performance evaluations.
Despite these concerns, the technology offers intriguing possibilities for academia and research. Traditional personality surveys are often constrained by small sample sizes and susceptibility to manipulation. By contrast, facially inferred traits enable the construction of large-scale datasets that could enhance our understanding of how personality shapes career trajectories and economic behaviour.
The ethical implications, however, loom large. Is it socially desirable, or even ethical, to make hiring decisions based on facial characteristics? The researchers acknowledge that while this technology could be programmed to deliver equitable outcomes across demographics, it may erode the principle of respecting individuality and effort.
As AI continues to infiltrate the labour market, policymakers and businesses will need to grapple with these questions. Can such tools be deployed in a way that augments fairness rather than exacerbates biases? Will they be used to complement traditional metrics, or will they supplant them altogether? The answers to these questions will shape the future of work in the AI era.
