The term “Artificial Intelligence” was coined by John McCarthy et al. in the mid-1950s, who defined it as “the science and engineering to develop intelligent machines that demonstrate critical thinking comparable to that of humans” (1). Since then, notable technological advances have improved the power and applications of AI tools, they have become an integral part of the personal and professional sphere.

In the field of medicine, AI systems have been developed to improve efficiency and precision, especially in tasks such as the diagnosis of pathologies from image data, and the processing of large volumes of clinical and diagnostic data, optimizing the provision of health care. Using machine learning algorithms, natural language processing, and computer vision, AI enables the analysis of complex medical data. (2).

An example of this is found in the imposing platforms that can be found today as part of the autoimmunity units in medium and high complexity laboratories, which make a great contribution to bacteriologists in this area who must carry out the reading and interpretation of complex techniques such as indirect immunofluorescence (IIF) in which, the accurate reading of a certain pattern on cells from specific cell cultures crucially directs the diagnosis and treatment of autoimmune diseases.

These software provide valuable support in the identification of specific characteristics that suggest an initial result with a certain level of confidence, where the reader is the one who authorizes or decides whether the reading generated by the platform that is fed by the constant and accurate classification of the user who directs it is approved.

This suggestion includes a standard and a titer (fluorescence level) for each patient. If the users’ classification is the same as that suggested by the platform, the software will feed its database, making the classification more and more accurate and providing a level of confidence very similar to that of the reader, whether this reading is accurate or not. However, it is important to highlight that the release of a result will always be under the responsibility of the reader, who must take into account correlating the patient’s medical history along with the other available requested tests. Other lines in which reading images in the laboratory has become very useful are the areas of hematology and cytohistology, providing greater agility to workflows.It is important to note that, despite the threats of AI regarding the replacement of jobs, in the clinical laboratory its scope is subject to the user because, although these systems have highly accurate algorithms, they are not infallible nor do they manage to have all the information that is possible in the laboratory or in the different services to give free rein to a result that does not fit the patient’s clinical condition.

This will always be a human responsibility and, although these platforms with AI are constantly feeding data, they will always learn from the user who executes them. If they do not have a solid base in the identification and classification of images in the specific area where AI knowledge is applied, it will learn erroneous concepts that will persist over time and will generate results that will not provide diagnostic value to patients.

Although AI continues to evolve, its implementation in health systems does not make it indubitable nor does it exceed the concepts and the correlation of highly qualified personnel in the jobs where it comes as an excellent help, it should be considered a complement and support to clinical laboratory professionals, cytohistologists and hematologists, contributing to optimize processing times, thus leaving more space to correlate highly complex patient data in the different health services.

This path still has a long way to go, legal and ethical issues to discuss, safer models to develop and implement in this area of health that must be approached with care, academia and discipline. It depends on us whether AI is our best ally in patient management or a threat that takes the place we have worked so hard to conquer. Different economic actors have already begun to receive the fruits of integrating AI into their business lines, and the health sector will not be the exception, but let’s not work for it, let’s always work with it seeking the benefit of our patients.

Bibliography

1. McCarthy J, Minsky ML, Rochester N, Shannon CE. A proposal for the Dartmouth summer research project on artificial intelligence, August 31, 1955. AI Mag. 2006;27:12–4

2. Bekbolatova M, Mayer J, Ong CW, Toma M. Transformative potential of AI in healthcare: definitions, applications, and navigating the ethical landscape and public perspectives. 2024;12:125. doi:10.3390/healthcare12020125