D’Agostino added that AI can be used, but decisions should always be made by a trained healthcare professional.
He recalled that AI use dates back to the 1960s and 1970s, as the so-called expert systems started off, but the main turning point was in 1980, when computers automatically carried out analyses based on large amounts of medical data.
In his statements, D’Agostino cleared up that in the 1990s the AI use in interpreting images such as X-rays or MRIs properly began in 1990s.
Then came another great movement, that of big data intelligence used to analyze large data sets that gave rise to what is possible today: prediction models, discoveries, and even drugs, he added.
This is not including the early detection of diseases from genetic data, and the most visible after the emergence of chatbots or tools to converse with machines: the big wave of the last few months with ChatGPT and other tools with which you can converse and from which you can ask for information.
D’Agostino considered that using AI it is possible to make diagnoses, but it must always be controlled or supervised by human beings.
“That is to say, do not let the AI makes the diagnosis and trust 100% in that diagnosis, because some failures have also been identified,” he said.
In this regard, D’Agostino insisted on the significance of using transparent or known algorithms, otherwise one does not know what is behind a software based on artificial intelligence. “If you don’t know what’s behind it, the risk is even greater,” he said.