A company builds an AI model to help screen job applicants by scoring resumes. During testing, the team notices that the model consistently gives lower scores to qualified candidates from certain demographic groups. Which responsible AI principle is most directly at risk and should guide the team's remediation?