Artificial intelligence will not replace doctors in clinical medicine but will serve as a tool to support healthcare professionals, expert radiologists told host Dr. Roberto Kalil on CNN Brasil.
Speaking on the television program CNN Sinais Vitais - Dr. Kalil Entrevista, scheduled to air on Saturday, August 8, at 7:30 p.m., the specialists emphasized that automated tools are designed to complement human care rather than eliminate medical positions.
The broadcast featured Giovanni Guido Cerri, a radiologist, full professor at the Faculty of Medicine of the University of Sao Paulo, and president of innovation center Inova HC, alongside Cesar Namura, radiologist and director of the radiology department at the Heart Institute, known as Incor. Both institutions operate within the Hospital das Clinicas complex in Sao Paulo, the largest public hospital network in Latin America.
According to the experts, artificial intelligence can support diagnostic decisions, reduce clinical errors, optimize hospital administrative processes, and increase overall health system efficiency. Namura noted that preventing medication dosage errors is a primary benefit, explaining that new drugs enter the market so frequently that healthcare staff administering them often cannot keep track of every potential drug interaction.
Radiology future and prediction of extinction
During the interview, Dr. Kalil joked that the specialty of radiology might eventually disappear as automated systems advance. Cerri responded by noting that a scientist had previously predicted the total extinction of the medical field due to technology, but that claim was ultimately proven false. Cerri explained that artificial intelligence should be viewed strictly as a decision support tool that helps clinicians select appropriate treatments.
Radiology relies heavily on digital imaging, including magnetic resonance imaging, computed tomography, and ultrasound scans, making it one of the medical fields most frequently linked to artificial intelligence development.
Algorithm flaws and population data
The guests cautioned that artificial intelligence systems can make mistakes, a limitation that medical professionals must keep in mind when interpreting recommendations. Namura explained that because humans program the software based on training data, poor initial data quality will inevitably produce flawed algorithms.
Namura added that artificial intelligence models are often validated on specific population groups that differ significantly from local hospital patients, which directly alters diagnostic accuracy. He advised clinicians using automated tools to examine where training data originated, how it was gathered, and what specific clinical question the system is being asked to solve.
The full interview will air on CNN Brasil on Saturday, August 8, at 7:30 p.m. local time.
