The amount of available medical knowledge now exceeds the organizing capacity of the human mind (2). Thus, we, as physicians, have started depending on the internet to supplement us with some of this information.
The reality of this web-enabled era is different (2). Artificial intelligence (AI) is a scientific discipline that focuses on understanding and creating computer algorithms that can perform tasks that are usually characteristics of a human’s cognitive abilities (4)
AI is being piloted in health care for faster and accurate diagnosis, to augment radiology, reduce errors due to human fatigue, decrease medical costs, assist and replace dull, repetitive, and labor-intensive tasks, minimally invasive surgery and reduce mortality rates. (4) In many cases, AI tools for medicine mostly play the role of a virtual assistant for physicians and healthcare systems, helping them to provide more accurate and efficient patient care. (1, 5) If time-consuming processes that require simple repetitive work are taken care of by AI, it would substantially reduce the fatigue of healthcare providers, and physicians could spend more time in facing with patients and concentrating on more complicated medical tasks. (5)
AI technology may also reduce the number of inadvertent errors in clinical practice and may decrease differences in judgments among medical professionals. (5)
Monitoring patient conditions 24 hours a day by AI systems, which would practically be impossible for humans to do, creates a safer patient environment. (5) Also, new patterns discovered by AI through the analysis of big data from clinical practice may lead to the development of new biomarkers for diagnosis and treatment. (5)
So how do we implement artificial intelligence? When using AI, one should acquire sufficient knowledge of basic and clinical medicines (which constitute the fundamentals of medical practice and are keys to understanding how to use AI for medicine), data science, biostatistics, and evidence-based medicine. (1) The skills required of practicing physicians will increasingly involve facility in collaborating with and managing artificial intelligence (AI) applications that aggregate vast amounts of data, generate diagnostic and treatment recommendations, and assign confidence ratings to those recommendations. (2)
Instead of attempting to decipher all of the facts of a case for themselves, physicians will have to know how to pose the right questions to machines, interpret their outputs, identify when machines make mistakes and course-correct accordingly, and communicate effectively with their medical colleagues and patients to formulate the best care plans. (3) The physician needs to understand the inputs and the algorithm and interpret the AI-proposed diagnosis to ensure no errors are made. (4) With machines at their sides, the best doctors won’t be the ones with the highest scores on standardized tests, but those with well-honed traits that machines cannot master, such as critical thinking, interpersonal and communication skills, emotional intelligence, and creativity. (3)
A new infrastructure for learning has to be introduced to help with using AI, and new educators from disciplines such as computer sciences, mathematics, ethnography, and economics will need to be hired. (4) AI training could be delivered via Continuing Medical Education (CME) programs and might need to be also taught by educators from outside the medical community. (4)
So what is the downfall? The development of AI algorithms almost as a rule requires data from a large number of patients. This raises many concerns including relating to patient privacy and control. (4) In situations where patient data are limited, algorithm developers train the models on synthetic or hypothetical data, with the risk of generating unsafe and incorrect treatment recommendations. (4) Also, although AI can curate and process more data such as medical records, genetic reports, pharmacy notes, and environment data and in turn retain, access, and analyze more medical information, it cannot replace the art of caring. (4)
We also have to keep in mind that the datasets used to train AI algorithms for medical applications are prone to various selection biases and may not adequately represent target populations in real-world clinical practice for many reasons (5)
AI systems are also vulnerable to cybersecurity attacks that could cause the algorithm to misclassify medical information. (4)
It is also unclear who is liable when a patient experiences serious harm because of an inaccurate prediction (4):the physician, the hospital, the company that developed the software, the person who developed the software, or even the person who delivered the data.
It is a new era in medical care. But the questions are still more than the answers. And even in the presence of technology like AI, we can never get rid of the human. So I or AI? I will always be involved to take care.
By Dr. Rola Ali-Hassan, CCFP
Consultant Family Medicine