This paper, forthcoming in the Handbook for Assessment in the Service of Learning, uses the Duolingo English Test (DET) as a case study to explore the critical role of Responsible AI (RAI) in high-stakes assessment. While AI offers efficiencies in item generation and scoring, it also presents risks, such as bias. The DET addresses these risks through four specific RAI standards—Validity and Reliability, Fairness, Privacy and Security, and Accountability and Transparency— demonstrating how these translate into practical implementation.