The application of generative pre-trained models in the medical field is driving the transformation of medical artificial intelligence (AI) towards a knowledge-driven paradigm. While their technical potential in auxiliary diagnosis, medical Q&A, and other scenarios has been widely verified, the high sensitivity of data and decisions in medical settings makes safety and compliance indispensable prerequisites for system deployment. This study aims to systematically identify the core challenges of medical generative pre-trained models in three dimensions: data privacy, regulatory compliance, and ethical governance, and construct a comprehensive governance framework with both theoretical support and practical feasibility. First, the research analyzes the risk transmission path of model privacy re-identification, model leakage, and harmful use from the perspective of technical mechanisms. Then, combined with the characteristics of medical data, it compares and analyzes the differences in compliance requirements under the HIPAA and GDPR frameworks, as well as the core pain points and solutions of technical adaptation. Subsequently, it sorts out the institutionalization trend of global medical AI ethical principles from soft initiatives to hard supervision, and proposes an ethical evaluation matrix covering four types of risks: cognitive, operational, social, and structural. Finally, it integrates institutional boundaries, technical boundaries, and ethical bottom lines to form a multi-level governance framework covering the entire life cycle of the model. The findings demonstrate that the sustainable development of medical generative pre-trained models critically depends on the construction of a “data–responsibility” trust chain, which urgently requires the coordinated evolution of technical solutions, institutional design, and ethical awareness. The core of future industry competition is not only the competition of algorithm performance, but also the systematic competition of governance capabilities and trust mechanisms.

Key Words: Medical GPT; data privacy; HIPAA; GDPR; AI ethics

Metaverse, a virtual reality world and human-computer exchange system created by integrating several digital technologies, is driving innovation and development in the healthcare industry and may bring disruptive shifts in the field of neurorehabilitation. This article reviews the research of virtual reality and brain-computer interface technology in the limb movement rehabilitation of stroke patients, in order to provide ideas for clinical rehabilitation treatment and basic research.

Key Words: metaverse; virtual reality; brain computer interface; stroke

With the acceleration of population aging and the increasing shortage of medical resources, artificial intelligence (AI) technology has shown great potential in the field of medical care. Integrating advanced technologies such as machine learning, natural language processing, and computer vision, AI has created innovative solutions in disease monitoring, rehabilitation assistance, elderly care, and mental health. The essence of AI care lies in the application of modern information technology and AI algorithms to provide intelligent support for disease prevention, monitoring, rehabilitation and daily health management. Its goal is to achieve roundthe clock monitoring and personalized management of patients by reducing labor costs and improving the efficiency of medical resource utilization. In this process, security, humanization, privacy protection, fairness and accessibility, as well as transparency and explainability, constitute the five basic principles of AI care, leading the development of technology in the right direction. In practical applications, AI monitors physiological indicators in real time through wearable devices to warn of health risks. Intelligent robots and VR technology provide customized guidance for rehabilitation patients; AI chatbots have become a new way of psychological comfort. At the same time, AI also plays an important role in optimizing care processes and improving administrative efficiency. However, the promotion of AI care will encounter multiple challenges such as technical obstacles, social acceptance, and policies and regulations, and it urgently needs the joint help of technological innovation, ethical guidance, policy adjustment, and multi-party collaboration. When promoting the application, it is necessary to establish clear data collection and analysis standards, manual review mechanisms, equipment security guarantees, and clear identification of AI identities. At the ethical and legal level, AI care needs to draw a clear line between technical and medical responsibility, ensure transparency of informed consent, and provide special protections for vulnerable groups. In addition, data quality, continuous learning, and multidisciplinary collaboration are also key factors driving the advancement of AI care technology. In the face of technical limitations, insufficient social acceptance, and lagging policies and regulations, the future development of AI care should focus on technological innovation, ethics education, policy improvement, and international cooperation.

Key Words: artificial intelligence; machine learning; natural language processing; healthy; care; ethics

Addressing the issue of resource scarcity for named entity recognition tasks in the medical field, a unified annotation methodology for special diseases entity corpora was formulated under the guidance of medical experts, and two special diseases entity corpora were constructed, namely Pediatric Bronchopneumonia Entity Corpus and Diabetes Entity Corpus. To verify the effectiveness of the proposed special disease entity corpus annotation method, the Pediatric Bronchopneumonia Entity Corpus was first compared with the publicly available dataset using BERT-BiLSTM-CRF and ERNIE-BiLSTM-CRF models. Then, the methodology was reapplied to diabetes electronic medical records to evaluate the robustness of the model. The results showed that both special diseases entity corpora got higher F1 scores than the public datasets, which suggests that special diseases entity corpus annotation methodology proposed in this paper has good robustness.

Key Words: electronic medical record; named entity recognition; corpus construction; Pediatric Bronchopneumonia Entity Corpus; Diabetes Entity Corpus

As an innovative teaching model, the Metaverse Teaching Rounds have realized the organic integration of advanced technologies such as virtual reality (VR), augmented reality (AR) and Medical Generative Pre-trained Transformer (MGPT), and built a highly immersive learning environment for students. This teaching mode can not only increase students’ interest in learning, but also help them better understand and master medical knowledge. The research progress of the metaverse empowering medical teaching rounds mainly includes the following fields: (1) The application of VR in teaching rounds. VR technology can improve students’ hands-on skills by providing them with a simulated medical environment and allowing them to practice in a virtual environment. (2) The application of AR in teaching ward rounds. AR technology can integrate virtual medical knowledge into the real medical environment, allowing students to learn virtual medical knowledge in the real medical environment, thereby improving their learning effect. (3) The application of MGPT in teaching ward rounds. MGPT is a natural language processing model based on deep learning technology, which can also be used for medical teaching rounds. At present, the world’s first digital human MGPT — BAIMGPT has been successfully developed and can be extended to teaching ward rounds. (4) By formulating scientific evaluation standards to evaluate the performance and harvest of teachers and students in the metaverse teaching rounds, improvements and optimizations can be made to improve and optimize the problems existing in the teaching process to improve the quality of teaching. In the future, the integration of VR, AR, and MGPT with advanced technologies such as mixed reality (MR) can build a more highly immersive learning environment for students, accelerate the integration of declarative knowledge and procedural knowledge, strengthen students’ theoretical foundation, improve their clinical thinking and practical ability, so as to better solve problems for patients.

Key Words: teaching rounds; virtual reality; augmented reality; mixed reality; artificial intelligence

Medical GPT, as a significant application of artificial intelligence technology in the healthcare field, has been explored in various areas, including medical imaging analysis, electronic medical record interpretation, disease prediction and diagnosis, and health management, demonstrating considerable potential for application. By using deep learning and natural language processing technologies, medical GPT can process and analyze vast amounts of medical literature and clinical data, thereby acquiring robust medical knowledge and reasoning capabilities. Current researches indicates that medical GPT has extensive application prospects in areas including intelligent diagnosis, health management, medical image analysis, drug research and optimization, and medical education and training. However, despite continuous technological advancements, the development of medical GPT still faces challenges in terms of data quality, privacy protection, security, and ethical regulations. Future development will require striking a balance between technological innovation and ethical regulations to ensure that medical GPT can evolve stably and healthily, bringing further innovation and value to the healthcare. 

Key Words: medical GPT; natural language processing; data security

With the rapid development of artificial intelligence technology, digital humans based on language models, such as GPT, have been widely applied in medical education. Digital human GPTs not only assist in learning medical knowledge and training clinical skills but also provide students with real clinical experiences through virtual patient simulations. Despite technical and ethical challenges, the personalized teaching and interactivity of digital human GPTs undoubtedly bring innovative changes to medical education. This article reviews the current literature to explore the application status, technological advancements, research findings, and challenges faced by digital human GPTs in the field of medical education， and looks forward to the future development directions of digital human GPTs in medical education.

Key Words: digital human GPT; medicine education

The registration of intellectual property rights in medical data is a crucial institutional arrangement for promoting the market-based allocation of data elements and advancing the development of metaverse medicine. Based on policy and legal foundations such as the “Twenty Data Articles” and considering the rapid growth trend of the global big data analytics market in healthcare, this paper conducts an in-depth analysis of the special legal requirements for medical data registration in terms of privacy protection and ethical review. By integrating three major typical cases from 2024 into the legal analysis, it systematically proposes legal pathways such as establishing a "medical data use right" and adopting a registration opposition system, providing theoretical support and practical references for constructing a medical data intellectual property registration system that meets practical needs.

Key Words: medical data; data elements; data property rights; trusted data space; medical data market