A groundbreaking study titled "AI Personality Extraction from Faces: Labor Market Implications" demonstrates that AI can extract the Big Five personality traits from a single facial image of professionals and predict their labor market outcomes such as MBA school ranking, starting salary, job seniority, career advancement, and job transitions even beyond traditional academic indicators like GPA or test scores .
How It Works: The "Photo Big Five" Model
Researchers from premier institutions (Wharton, Yale, Indiana) used an algorithm based on Kachur et al. (2020), which was pre-trained on annotated images with self-reported personality scores (Big Five: Openness, Conscientiousness, Extraversion, Agreeableness, Neuroticism). The algorithm extracts subtle facial cues to infer personality with correlations between 0.14 and 0.36 compared to self-reports .
Mechanism Behind the Extracted Traits
Genetic factors & facial symmetry: Up to 60% of personality traits may be genetically linked to facial structure.
Environmental influences: Hormonal exposure impacts both facial development and personality.
Feedback effect: Facial appearance influences how we behave and how others perceive us, reinforcing certain trait expressions .
Major Findings
1. Predicting Reputation and Compensation
The “Photo Big Five” score predicts MBA school rank, initial salary, career growth over time, and job switching behavior .
Movement from the bottom to the top quintile of “desirable” Photo Big Five traits correlates with an 8–12% increase in compensation for both men and women and 2–3% more annual compensation growth .
2. Adds Insight Beyond Academics
The facially derived personality traits are only modestly correlated with GPA and standardized tests but add nearly equivalent additional predictive value for salary outcomes .
3. Comparable to Traditional Bias Factors
Photo Big Five scores predict outcomes as strongly as race, attractiveness, and educational background. For men, the compensation gap between top and bottom personality quintiles exceeds racial pay disparities. For women, it accounts for ~65% of the Black-White compensation gap .
4. Broader Implications
This methodology can scale personality assessment via public images. While it delivers powerful predictive insight, it also raises serious ethical concerns potential privacy violations, risk of discrimination, and undermining fairness and autonomy .
Ethical Considerations
Discrimination based on immutable traits: Screening using facially predicted personality could unfairly disadvantage people based on features they cannot change .
Transparency deficit: Automatically aligning personality with career potential may entrench systemic biases, especially in marginalized groups.
Lack of oversight: The model treats personality as objective data, but black‑box neural networks could encode biases without checks .
Why This Isn’t Just Sci-fi
Large empirical basis: Study uses over 96,000 MBA professionals’ LinkedIn photos and administrative salary/job records
Observed real-world effects: Statistical analysis shows consistent correlations across multiple cohorts and genders not spurious artifacts .
Conclusion
The Photo Big Five model illustrates AI's capacity to decode personality from facial features and use it to predict real labor market success. While technologically impressive and potentially revolutionary for scaling personality insight, its deployment in hiring or admissions demands urgent ethical governance to prevent bias, invasion of privacy, and unfair screening.
Further Reading & References:
Guenzel, Kogan, Niessner & Shue (2024). “AI Personality Extraction from Faces: Labor Market Implications” (SSRN)
“AI can predict career success from a facial image” (Computerworld)
Yale SOM insights on AI photo-personality and careers