AUTHORS: Shane O’SullivanSimon LeonardAndreas HolzingerColin AllenFiorella BattagliaNathalie NevejansFijs W.B. van LeeuwenMohammed Imran SajidMichael FriebeHutan AshrafianHelmut HeinsenDominic WichmannMargaret Hartnett Via



The application of human‐centred artificial intelligence (AI) and transparent machine learning (ML), to integrated Gross anatomy models, complemented with medical imaging data of cadavers and novel audio tissue characterization methods, makes available a route to autonomous robot‐delivered surgeries and pathological characterisation. It can provide the capability of objective autopsy with reliable operations that improve both accuracy (robot) and tissue interpretation (AI), thus preventing the wrong diagnosis from being made.


We reviewed technological advances and state‐of‐the‐art developments documented by undertaking a literature search on autonomous robotics for surgery and autopsy, tracing agents, explainable AI, ML black box solutions, algorithmic transparent/opaque processes, as well as AI legal and ethical issues such as data biases (e.g. gender/racial/social bias) and relevant traditional, religious or sociocultural aspects. For the use of autonomous robotics in either surgery or autopsy, our approach is to discuss the ‘challenges and knowledge gaps’ followed by our proposed ‘hypotheses and recommendations’.


The integration of explainable AI and ML, and novel tissue characterization sensorics to tele‐operated robotic procedures with medical imaged cadavers, provides robotic guidance and refines tissue classifications at a molecular level. This can lead to the development of software that is not restricted to just automating ‘autopsies’ but, rather, could hold potential for ‘surgeries’ at large. The current R&D on autonomous surgical robotics is typically presented by comparing the skills of ‘humans’ versus ‘robots’. In contrast, our assertion is that autonomous robotics and explainable AI can support healthcare values complementing and augmenting human capacities – NOT replacing them. Autonomous robots, trained to perform surgery or autopsy, based on their trained algorithms, can deliver practical positive solutions in medicine. This potential improves when advanced imaging solutions help provide GPS‐like guidance and improve tissue interpretation. In the short term, we are confident that autonomous robotic ‘autopsy’ coupled to ML, with a human‐in‐the‐loop, could present a fundamental change in forensic analysis of cadavers. Furthermore, autonomous procedures, applied to medical imaging‐guided robotic autopsies, can furnish powerful support to pathological and technological developments that are translatable to broader surgical procedures (e.g. enhancing motor control in a standard tele‐operated surgical system and improving imaging based GPS‐like navigation of robotic tools). In addition to being necessary for complex access and guidance, these reliable operations are independent to the human factors that encompass fatigue, and risk of pathological exposure, such as contagious vectors and psychopathy.

O’Sullivan, S, Leonard, S, Holzinger, A, et al. Anatomy 101 for AI‐driven robotics: Explanatory, ethical and legal frameworks for development of cadaveric skills training standards in autonomous robotic surgery/autopsy. Int J Med Robotics Comput Assist Surg. 2019;e2020.

Full text to be added soon.