Sechenov University: Method of applying artificial intelligence to train future doctors

The main articles are:

• Speech Recognition (Technology, Market)
• Speech technology: On the path from recognition to understanding
• Artificial Intelligence in Medicine

2024: Representation of the method of applying AI in education, based on a socratic approach

Young scientists at the First Moscow State Medical University have developed a method for applying AI in education based on a socratic approach. With the help of a certain sequence of questions that AI asks on a virtual microdrug (a thin section of tissue placed on a slide for consideration under a microscope), students independently learn to recognize various pathological processes and diagnose diseases. This method will make it possible for future doctors to study morphological disciplines and prepare for exams in the most convenient chat format, and for teachers to allocate more time to analyze complex clinical cases with students. The university announced this on December 24, 2024.

The Socrates method, that is, a form of reasoned dialogue based on questions and answers, may become a new standard for the use of artificial intelligence in education, says the author of the project, a 4-year student, a research trainee at the Borovsky Institute of Dentistry of the First Moscow State Medical University, general director of the university startup "Architects" Daria Archakova. The method does not require expensive additional training of models. To use it, teachers just need to create a reference knowledge base on a narrow topic, and programmers to develop an architecture in which the language model refers to verified information and asks unique questions about the student's microdrug. In the future, this will automate education, freeing up more time for teachers to personally train future doctors.

Before us, many development teams have tried to train artificial intelligence to answer students' questions. But generative AI hallucinates and tends to produce deliberately false answers. Therefore, it has not yet been introduced into the educational process, especially in the training of future doctors. Our method has circumvented these limitations and put artificial intelligence in the service of students and teachers.

The reference collection of the Institute of Clinical Morphology and Digital Pathology, which AI uses to conduct dialogue with students, includes more than 200 virtual micropreparations in norm and pathology, and more than 2 thousand annotations to them. All shots are available in ultra-high quality. Ultra HD 4K By the end of 2024, the project team plans to increase the collection of digital images of microdrugs to 500.

AI learning technology looks like this: a student studies monitor computer tablet a microdrug in a digital image or digital image. His assistant - artificial intelligence - asks questions on the virtual slide in a certain order. Consistently responding to them, the student develops clinical thinking and learns to diagnose diseases.

Digital technologies and artificial intelligence are no longer the future, this is the reality of modern medicine, modern healthcare, and in this regard, education, educational programs, teaching methods are changing, - said the scientific director of the project, Vice-Rector for Scientific Activities of Sechenov University, Director of the Institute of Clinical Morphology and Digital Pathology, Doctor of Medical Sciences, Professor Tatyana Demura. - We actively use electronic tutorials, as well as developments such as "Virtual Patient." In relation to morphological disciplines, in addition to theoretical knowledge, it is also necessary to master practical skills - to be able to navigate in microdrugs, in histoscans. With the help of artificial intelligence technologies, students can now learn to recognize various pathological processes in them.

Already at the beginning of 2025, this method of using AI, which forms clinical thinking using the Socratic dialogue and deep survey of students, will be tested on the educational platform of the Institute of Clinical Morphology and Digital Pathology of Sechenov University. The results of the scientific study will be presented at an international educational conference in 2025.