AI Recruiting Tool Systematically Excluded Veterans Due to Employment Gap Bias

In early 2025, multiple companies using AI-powered recruiting tools were found to be systematically screening out qualified military veterans from job applications. The discrimination was discovered through analysis by veteran advocacy groups who noticed significantly lower callback rates for veterans compared to civilians with similar qualifications. The AI systems were interpreting periods of military service as employment gaps, treating them as negative indicators in candidate scoring algorithms. The bias manifested in several ways: first, the algorithms flagged military service periods as unexplained employment gaps, automatically reducing veteran candidates' scores. Second, the systems failed to recognize military occupational specialties (MOS) and their civilian equivalents, treating military job titles as irrelevant or unknown experience. Third, the AI tools penalized non-linear career progression patterns common among veterans transitioning to civilian careers. Veteran advocacy organizations, including Iraq and Afghanistan Veterans of America (IAVA) and Veterans of Foreign Wars (VFW), filed complaints with the Equal Employment Opportunity Commission after conducting their own testing. They submitted nearly identical resumes with only military service versus civilian employment during the same periods and documented significant disparities in advancement rates through AI screening systems. The EEOC launched an investigation in January 2025, issuing guidance clarifying that AI hiring tools must comply with existing protections for veterans under the Vietnam Era Veterans' Readjustment Assistance Act and Americans with Disabilities Act. The Commission noted that while not intentional, the systematic exclusion of veterans constituted disparate impact discrimination requiring immediate remediation. Several major corporations using these AI recruiting platforms acknowledged the issue and began implementing corrective measures, including retraining algorithms with veteran-inclusive datasets and adding human review checkpoints for veteran applications. However, the incident highlighted the broader challenge of ensuring AI fairness for protected classes with non-traditional employment patterns.