A financial services company builds an AI model to recommend which customers should be offered premium credit products. During testing, the team notices that qualified applicants from certain neighborhoods are consistently scored lower than equally qualified applicants from other areas. Which responsible AI principle is most directly at risk and should guide the team's corrective action?