A large-scale TCM model integrated into hospital workflows that enhances clinical practice by providing AI-assisted support throughout the hospital's workflow.
GuangYi QiZhi
Scholarly TCM bot trained to interpret classical Chinese medical literature that supports academic research and historical text interpretation.
Rule-based expert systems that mimicked TCM diagnostic logic. Early prototypes were used in hospitals and universities for clinical simulation.
TCM Diagnostic Expert Systems (Legacy Bots)
Collaboratively developed by Chinese institutions, it simulates clinical consultations and diagnostic reasoning using TCM principles. It features syndrome differentiation, herbal and acupuncture recommendations, and patient education.
Trained on classical texts and TCM clinical guidelines, it interprets symptoms and provides syndrome analysis and treatment suggestions. It is commonly used in medical education, diagnostic support, and case simulation.
TCM Robotics
About Robotics
Robotics is a field of engineering and computer science concerned with designing, constructing, operating, and using automated machines capable of performing tasks traditionally carried out by humans. Robotics integrates hardware (mechanical components), software (algorithms and control systems), and often Artificial Intelligence (AI) to enable machines to perceive, decide, and act in dynamic environments.
In traditional Chinese medicine (TCM), robotics, when combined with AI, represents an emerging frontier in modernising clinical practice. Robotic systems are not merely mechanical tools; they are becoming intelligent collaborators capable of supporting precision, consistency, and efficiency in diagnosis and treatment.
Advantages of Robotics in AI-TCM Integration
-
Standardisation: Robots eliminate variability caused by human fatigue or subjectivity.
-
Precision: Enhanced ability to perform delicate tasks (e.g. needling, sensing pulse) with micrometre accuracy.
-
Safety: Robots can detect anomalies or errors and stop procedures before harm occurs.
-
Training: Robotic systems can simulate patient interaction for educational purposes, helping students refine diagnostic and therapeutic skills in controlled environments.
Limitations and Considerations
-
Lack of Intuition: Robots cannot yet replicate the intuitive and holistic reasoning of experienced practitioners.
-
Ethical Concerns: When robots are involved in care, questions about patient trust, consent, and human connection arise.
-
Cost and Accessibility: High costs may limit widespread adoption, particularly in rural or traditional settings.
When combined with AI, robotics offers TCM a powerful means of enhancing diagnostic accuracy, procedural safety, and educational quality. While it will not replace TCM practice's intuitive, human-centred nature, it is a valuable partner in the digital transformation field, bridging ancient wisdom with modern technological capability.
TCM Bots
TCM bots are intelligent software applications—often powered by Artificial Intelligence (AI) and Natural Language Processing (NLP)—designed to interact with users, provide Traditional Chinese Medicine (TCM) knowledge, and assist with diagnosis, treatment guidance, and education. These bots simulate human conversation and clinical reasoning, making TCM more accessible, interactive, and standardised. Types of TCM Bots and Their Functions include:
-
Diagnostic Assistants: These bots ask patients questions based on the Four Diagnostic Methods (inspection, listening/smelling, inquiry, and palpation), then use AI models to suggest possible TCM syndromes, recommend acupuncture points or herbal formulas, and provide lifestyle or dietary advice.
-
Educational Bots: Used by students and practitioners to quiz users on TCM theory (e.g. meridians, Five Elements, Zang-Fu organs), simulate case studies and diagnosis and explain classical texts and formulas (e.g. from Huangdi Neijing)
-
Patient-Facing Chatbots: These help general users understand their constitution (e.g. Qi deficiency, Yin excess), self-care strategies rooted in TCM and preventive healthcare tips.
-
Pharmacy and Formula Bots: These assist in herbal formula recommendation and adjustment, cross-checking herb compatibility and identifying contraindications or herb-drug interactions.
Many bots use advanced models such as ChatGPT, TCM-GPT, or QiBo, which are pretrained on large TCM corpora including classical texts, clinical records, and modern research.
TCM bots represent a fusion of ancient medical wisdom and cutting-edge AI. They act as digital assistants, educators, and advisors—helping modernise TCM while preserving its core diagnostic philosophy.
It is designed for TCM theoretical frameworks and focuses on syndrome classification and diagnostic simulation.
Self-assessment tools guide users through a consultation-style interaction that features constitution typing, dietary advice, and herbal suggestions.
Built on ChatGPT with added acupuncture and TCM datasets, it offers point selection, diagnosis guidance, and treatment planning.
Developed by the Chinese academic and AI research community, it is a multimodal AI model that integrates TCM clinical, visual, and text data for intelligent assistance.
Developed by Chinese medical informatics groups, they are specialised TCM chatbots used for data-driven diagnosis and herbal compatibility checks.
Provides conversational support for patient questions, constitution typing, and herbal suggestions.
ChatTCM (Experimental)
A TCM-specific large language model (LLM) designed for expert-level medical diagnosis and syndrome differentiation-based treatment. Supports practitioners in delivering precise diagnoses and personalised treatment plans.
A TCM-focused LLM was developed to address challenges in TCM diagnosis and syndrome differentiation that enhance diagnostic and differentiation capabilities, offering valuable insights for future research.
A healthcare chatbot for TCM applications, integrating 3D human body visualisation. It assists users in understanding TCM concepts and therapies through interactive visualisations.
An AI-powered TCM consultation tool designed for experienced practitioners that supports practitioners in refining their practice with detailed diagnostic and treatment support
A specialised chatbot designed to support TCM students and practitioners in learning that enhances the learning experience by facilitating the understanding of complex TCM concepts.
TCM Study Buddy (YesChat)
An AI-powered mini-program for tongue diagnosis and dietary guidance that Streamlines TCM consultations by integrating traditional tongue diagnosis with AI image recognition technology
Wang She Wen Shan
Robotic-Assisted Acupuncture
Robotic acupuncture is an emerging frontier in integrating artificial intelligence (AI), robotics, and Traditional Chinese Medicine (TCM). It seeks to enhance the delivery of acupuncture by leveraging advanced technologies to achieve greater precision, consistency, and potential accessibility, particularly in remote or clinical settings where traditional hands-on treatment is not feasible.
Robotic-assisted acupuncture involves using intelligent systems to identify, insert, and stimulate acupuncture points on the human body, autonomously or under a trained practitioner's supervision. These systems are designed not to replace human acupuncturists but to serve as clinical assistants or extend care capabilities, particularly in scenarios requiring high accuracy or telemedicine solutions. Key Features of Robotic-Assisted Acupuncture include:
-
3D Anatomical Mapping and Point Localisation: Advanced imaging technologies, such as 3D body scanning and anatomical landmark detection, allow robotic systems to precisely locate acupuncture points based on the patient’s unique morphology. AI algorithms customise point identification beyond static textbook positions, enabling tailored treatment for each individual.
-
Robotic Needle Insertion and Manipulation: Precision robotic arms are engineered to control needle insertion with exact depth, angle, and force. Some systems can even simulate traditional manual techniques, such as lifting, thrusting, or twirling, ensuring therapeutic accuracy and consistency.
-
Sensor Feedback and Real-Time Adjustment: Built-in pressure, resistance, and biofeedback sensors (e.g., EMG, GSR) monitor the patient’s response during needling. These allow the system to adjust stimulation dynamically, optimising comfort and effectiveness while maintaining safety.
-
Non-Invasive Alternatives: In jurisdictions with regulatory or ethical concerns regarding automated needling, robotic platforms are being developed for laser acupuncture or electroacupuncture, offering non-invasive solutions with automated precision in point stimulation.
-
Teleoperated and Remote-Control Capabilities: With growing demand for accessible healthcare, robotic acupuncture systems offer the potential for remote treatment. Licensed practitioners can supervise or operate robotic devices from a distance, delivering care to underserved or geographically isolated populations.
China has led efforts in developing robot-assisted acupuncture systems for treating musculoskeletal and neurological conditions, often in research hospitals and university settings. Japan and Korea have piloted robotic arms for guided needling using electromyography (EMG) feedback to increase accuracy. The United States and Europe have focused primarily on robotic laser acupuncture and electrostimulation platforms, prioritising non-invasive applications for safety and accessibility. Future applications may include fully autonomous systems in space medicine, eldercare facilities, or even robotic clinics where AI-based diagnostic tools (e.g., facial recognition, pulse analysis, tongue scanning) are integrated into one unified treatment platform.
Robotic acupuncture illustrates a transformative path forward where TCM can engage with high-precision technology to maintain its therapeutic depth while expanding its reach and adaptability. As these technologies evolve, they promise to uphold the art of acupuncture within the framework of modern healthcare innovation.
China, 2025
Utilises the OptiTrack motion capture system for precise acupoint localisation, a collaborative robotic arm for needle insertion, and a De Qi sensation detection system. Achieved a positioning error within 3.3 mm, aligning closely with professional acupuncturists' standards. Needle insertion trajectory maintained a mean deviation distance of 0.02 mm and a deviation angle of less than 0.15°. The De Qi detection system demonstrated an average accuracy of 95.19%, effectively identifying sensations like "Tong," "Suan," "Zhang," and "Ma. Designed to streamline the acupuncture process, increase efficiency, and reduce practitioner fatigue, with potential applications in treating conditions like low back pain.
OptiTrack-Based Acupuncture Robot
China–USA, 2025
Combines ultrasonic computed tomography (UCT) with mixed reality (MR) to visualise acupuncture points in real-time. Provides a 3D anatomical reconstruction of the patient's arm, allowing for accurate needle insertion guided by MR overlays. Employs an attention-adaptive 3D user interface (3DUI) to assist practitioners during needle placement. Aims to enhance training for new practitioners and medical students by providing visual aids and real-time guidance, potentially reducing the learning curve associated with mastering acupuncture techniques.
MRUCT: Mixed Reality Acupuncture Guidance System
China, 2025
A hybrid model combining the Mamba state-space model and Transformer architecture to achieve efficient global information integration for acupoint localisation tasks. Achieved an average Euclidean distance pixel error (EPE) of 7.792 on a private dataset of acupoints on the human back. Demonstrated an average time consumption of 10.05 milliseconds per localisation task, improving accuracy and speed by approximately 14% compared to previous models. Enhances the efficacy of acupuncture treatment and demonstrates the commercial potential of automated acupuncture robot systems.
RT-DEMT: Real-Time Acupoint Detection Model
China, 2024
Develop an acupuncture robot capable of replacing manual treatments for specific points. Utilises machine vision for spatial registration of the robotic arm, patient, acupuncture tools, and optical navigator to locate acupuncture points accurately. Implements a needle insertion mechanism guided by a trajectory algorithm that ensures precise needle placement while avoiding obstacles. Aims to standardise acupuncture treatments and improve the reproducibility of needle insertion techniques.
Analysis and Design of an Acupuncture Robot System
AI-Powered Tuina Massage Robots
Tuina massage is a traditional hands-on therapeutic technique rooted in Traditional Chinese Medicine (TCM) principles. It is primarily used to promote the flow of qi and blood through the meridians, restore physiological balance, and support immune function. Tuina serves preventive and curative functions, treating various musculoskeletal and internal disorders by manipulating the body's soft tissues and energy pathways.
Integrating Artificial Intelligence (AI) into tuina massage therapy represents a frontier in modernising this ancient practice. Although still in its developmental stages, AI-powered massage robotics is emerging as a promising tool for enhancing clinical efficacy, increasing accessibility, and improving treatment standardisation. These intelligent systems aim to replicate core tuina techniques such as acupressure, vibration, and percussion. However, current technologies face challenges in fully reproducing the complexity and nuance of manual methods such as gentle stroking, dynamic stretching, and intricate kneading.
Two primary categories of robotic tuina technology are under development: vibration-based and percussion-based massage systems. These devices utilise AI algorithms and sensor technologies to simulate acupressure and deliver consistent pressure to specific points, providing a semi-automated version of tuina therapy. While they cannot yet replicate the dexterity and intuition of a human practitioner, they offer repeatable, data-driven treatments. They can be especially useful in rehabilitation, home care, and wellness settings.
EMMA (Expert Manipulative Massage Automation) is a notable innovation in this space, developed by AiTreat Pte Ltd in Singapore. EMMA is an AI-driven therapeutic massage robot designed to deliver precision tuina massage. It uses advanced sensor arrays, including pressure sensors and stiffness detectors, to locate areas of muscle tension and acupoint regions. EMMA can automatically adjust the intensity and angle of treatment in real time, based on the patient's feedback and biomechanical data. Key features of EMMA include:
-
Machine learning algorithms that process muscle resistance patterns and physiological feedback to personalise massage protocols.
-
Internet of Things (IoT) connectivity, enabling EMMA to store and analyse treatment data over time for ongoing refinement.
-
Acupoint recognition models that allow EMMA to simulate TCM diagnosis by identifying therapeutic zones through sensor data, mimicking the pattern-recognition skills of an experienced practitioner.
These capabilities make EMMA more than a mechanical tool—it functions as an intelligent therapeutic assistant. It bridges the diagnostic and treatment processes, using AI to determine effective intervention zones and modulate technique dynamically. Through continued training and data collection, such AI systems can improve in precision and efficacy over time, potentially supporting the work of human practitioners and expanding access to therapeutic massage care.
As research and development in this area progress, AI-powered tuina robots will likely become increasingly sophisticated. Future advancements may include multi-modal systems integrating facial recognition, thermal imaging, and biometric feedback to create more personalised treatment experiences. Ultimately, AI-powered tuina massage represents a key example of how technology can extend the reach and capability of TCM, ensuring that its time-tested principles continue to benefit patients in the modern era.