Mental health disorders create profound personal and societal burdens, yet conventional diagnostics are resource-intensive and limit accessibility. Recent advances in artificial intelligence, particularly natural language processing and multimodal methods, offer promise for detecting and addressing mental disorders. However, these innovations also introduce privacy concerns. Here we examine these challenges and propose solutions, including anonymization, synthetic data and privacy-preserving training, while outlining frameworks for privacy-utility trade-offs, aiming to advance reliable, privacy-aware artificial-intelligence tools that support clinical decision-making and improve mental health outcomes.