The field of mental healthcare is undergoing significant developments with the integration of large language models (LLMs) and multi-agent systems. One of the key directions is the automation of clinical interviews and assessments, with a focus on improving accessibility and accuracy. However, several challenges have been identified, including the expression of stigma and inappropriate responses by LLMs, as well as their limited ability to replicate human-like therapeutic relationships. To address these challenges, researchers are exploring the development of more advanced LLM-based dialogue systems that can conduct formal diagnostic interviews and assessments. These systems aim to bridge the gap in mental healthcare accessibility by providing a more structured and clinically rigorous approach. Noteworthy papers in this area include: MAGI, a multi-agent guided interview framework that combines clinical rigor, conversational adaptability, and explainable reasoning. TRUST, an LLM-based dialogue system that replicates clinician behavior and performs comparably to real-life clinical interviews. These developments highlight the potential of LLMs and multi-agent systems to transform the field of mental healthcare, but also emphasize the need for further research and evaluation to ensure the safe and effective deployment of these technologies.