Artificial Intelligence (AI)

A Little Background

AI is a broad field of computer science that deals with the creation of intelligent machines that are capable to simulate human-like intelligence to undertake cognitive tasks such as problem-solving, and decision-making but more efficiently. The concept of AI has been around for many decades and can be divided into several subfields, which include Machine Learning (ML), Natural Language Processing (NLP), robotics and automation. A specific type of AI referred to as an “expert system” focuses on decision-making and has been used for some time in the fields of medicine, law and finance amongst others. These expert systems use rules and large knowledge bases to make decisions. One of the most famous examples of such a system in healthcare was the medical diagnosis system MYCIN, which was developed at Stanford University about 50 years ago to aid in the diagnosis and treatment of bacterial infections. More recent clinical decision support systems (CDSS) are able to provide doctors with real-time guidance for patient diagnosis, treatment plans, and drug interactions. AI is also used for predictive modelling, the ability to predict the likelihood of future outcomes based on historical data, which is particularly useful for at-risk patients.

In more recent decades as algorithms became more powerful coupled with the increase in data availability, AI development has made rapid and significant progress in the area of “deep learning,”. A sub-field of ML, deep learning involves training multi-layered neural networks and is used to improve the accuracy of medical imaging (CT and MRI ), as well as analyse huge amounts of medical data, e.g. electronic health records. From these advances, there have also been significant developments in another sub-field of AI, known as natural language processing (NLP). These systems can understand and generate human-like text. Google is reliant on NLP algorithms. The capabilities of NLP are currently being seen through OpenAI’s ChatGPT, Google’s Bard, and Microsoft’s Bing Chat, which are the most advanced to date.

Why do future doctors need to learn about AI?

To be prepared for health care in the digital age requires understanding the transformative effect that AI will have on the profession. This means not only having an understanding of technology’s likely impact but also how to utilise such technologies competently. This involves understanding how AI is used and how to interpret its results, and how it makes predictions and recommendations to support its clinical decision-making. Additionally, it is crucial to be aware of the limitations of AI, such as patient privacy concerns, data security issues, and the possibility of bias and discrimination along with the ethics and regulations that govern the use of AI in healthcare.

NLP techniques are also likely to play a major role in health care, such systems can already be used to provide efficiencies for example, extracting information from unstructured clinical notes, medical literature and other sources, and assisting with diagnosis and treatment planning. Although, the overhead in developing such capabilities requires considerable effort commencing from the collection of a huge amount of good quality data (perhaps from EHR systems). After which this data needs to be cleaned and reformatted before being annotated by medical experts before it can be used to train the NLP system. And it will need to be evaluated again by experts.