A software company is building a voice-controlled banking assistant. During design reviews, the team decides to add support for multiple languages, screen-reader compatibility, and alternative text-based input for users who cannot speak clearly or at all. Which responsible AI principle is the team primarily applying with these design decisions?