Artificial intelligence (AI) is a wonder and a gift of technology to mankind. It is rapidly transforming everything from each field. It is also transforming mental health care by enabling scalable screening, personalized interventions, symptom monitoring, and task automation. AI-driven conversational agents and algorithmic screening tools show promise in expanding access to care and delivering evidence-based interventions such as cognitive behavioural therapy (CBT) at scale. However, variable evidence quality, safety concerns (including crisis detection and harm amplification), privacy risks, algorithmic bias, and regulatory gaps present substantial challenges. This paper reviews current applications and evidence, summarizes ethical and regulatory considerations, and proposes recommendations for safe, effective implementation of AI-powered mental health supports. Key recommendations include rigorous clinical assessment, human-in-the-loop designs, transparent reporting, robust privacy safeguards, and regulatory alignment with digital health guidance.