Reflection on “Health Data Science, AI and Precision Medicine”

By Yuen Ching Yu, Holy Trinity College

Starting from defining the health data science and introducing various health data types, followed up by applications of health data, and ending by talking about the use of AI, the lecture has greatly expanded my knowledge in health data and AI. I have also learnt about what precision medicine is. Aside from that, having been taught the uses of health sensors and health data in various fields like predicting diseases, detecting anomalies, recognizing sleeping patterns, etc. I was astonished by how helpful and beneficial it is in using those sensors in medical fields. Moreover, elaborating on the application of AI to Covid-19 detection, allows me to understand how AI is used in disease treatments more clearly. I like how Dr. Li relates AI to our daily lives, it has drawn my attention. When talking about medical imaging data, Dr. Li asked if we can figure out which photo is real and which is generated by computer algorithms, which makes the lecture fun and interactive. I also enjoyed a lot in the Q&A session which we as an audience could finally get a chance to ask questions and clear misconceptions. To conclude, having been given a passionate and engrossing lecture, I have found the knowledge taught by Dr Li fascinating and helpful as I kept focusing throughout the whole lecture. Besides, I have been captivated to explore more on those topics.

Through my investigations online, I have found that experts have said that the healthcare jobs most likely to be automated, in other words being replaced by AI, would be those that involve dealing with digital information, radiology and pathology for example, instead of those with direct patient contact. Despite the significant advantages brought by AI, AI is said to be unlikely to bring about a substantial change in healthcare over the next 20 years or so, as there are tasks that only humans can do and AI cannot do. For example, radiologists are unlikely to be replaced right now. Radiologists do more than just read and interpret images. Like other AI systems, radiology AI systems can only perform single tasks, similarly deep learning models in labs and startups are only trained for specific image recognition tasks. However, thousands of such narrow detection tasks are necessary to fully identify all potential findings in medical images, and only a few of these can be done by AI today. Not only does AI cannot fully analyze and apply related findings to treatments, AI is also unable to consult with other physicians on diagnosis and treatment like radiologists. Therefore, it is said that the penetration of AI into some healthcare fields like radiologists is likely to be slow.

To sum up, there are limitations in AI. Nonetheless, with a longer and deeper investigation on AI, I believe that AI will have an important role to play in the healthcare offerings of the future. In the form of machine learning, the primary capability behind the development of precision medicine will be sorely needed advance in healthcare. The challenges faced by AI will ultimately be overcome, taking a long time to do so, and the use of AI as well as data science in the medical field will be more and more mature. I will look forward to the development of AI.