The use of artificial intelligence (AI) in medicine has been on the rise, but there are concerns that over-reliance on this technology may lead to a loss of clinical knowledge. In a viewpoint published in JAMA, Agnes Fogo, MD, and colleagues highlighted the potential consequences of relying too heavily on AI in the field of pathology. They used the example of kidney pathology, where pathologists are now utilizing AI algorithms to annotate tissue specimens. While this has the potential to improve efficiency and streamline the diagnostic process, it also comes with risks.
In the field of kidney pathology, it is estimated that up to 100,000 annotations are needed before an AI algorithm can recognize basic subunits such as a glomerulus. Once the algorithm has been trained, it can quickly identify these elements, making the pathologist’s job easier. However, the downside is that the pathologist’s skills in evaluating basic histology elements may gradually diminish over time. With the basic elements being moved out of their daily practice, the real intelligence and understanding of the kidney’s architecture may be lost.
Another concern highlighted by Fogo and colleagues is the potential loss of problem-solving skills. If AI models are relied upon too heavily to streamline or innovate in medicine, the capacity for physicians to understand and solve problems without the help of AI may diminish. The authors point out that advancements in AI may lead to the identification of new tissue areas that were not previously named in traditional pathology. However, if little effort is made to understand these newly identified areas, the knowledge of disease mechanisms underlying their development may be lost.
One of the major challenges posed by AI in medicine is the issue of “black box computing.” This refers to the inability to understand how an AI arrives at its conclusions. Fogo and colleagues warn that if medicine reaches a stage where the output is defined in a black box that is no longer consistent with previously defined entities, much of the knowledge on disease mechanisms will be forgotten. This could result in a situation where healthcare decisions are solely based on intervention strategies without a deep understanding of the underlying disease processes.
Striking a Balance
While there are valid concerns surrounding the over-reliance on AI in medicine, it is important to acknowledge its potential benefits. AI models have shown great potential as support tools in clinical decision-making, improving efficiency and accuracy. Fogo and colleagues emphasize the need for physicians to contemplate how to harness the potential benefits of AI without sacrificing control over their profession.
Physicians should recognize that keeping AI within boundaries is essential for the survival of their profession and for meaningful progress in diagnosis and understanding disease mechanisms. While AI can be a valuable tool, it should not replace clinical knowledge and expertise. It is crucial for doctors to maintain their skills and continue to actively engage with the interpretation of diagnostic findings, even when aided by AI algorithms.
The over-reliance on AI in medicine poses risks to the field of pathology and clinical knowledge. While AI can streamline the diagnostic process and improve efficiency, it may also lead to a loss of skills in basic histology elements and problem-solving capacities. Striking a balance between utilizing AI as a support tool and maintaining clinical expertise is crucial for the progress of medicine. Physicians should remain vigilant in ensuring that AI remains within boundaries, allowing for meaningful progress and the preservation of their profession.
Leave a Reply