Fraunhofer AICOS researchers are coordinating a project that aims to create mechanisms, through artificial intelligence, so that the issue of privacy is not an obstacle to sharing health data. The exchange of information allows to protect patients and also to improve research. However, this exchange has not always been possible.
“Each hospital or each service has its own database and the others cannot have access” to certain information, he explains to TSF the researcher David Belo, who justifies it with the existence of “very restricted privacy issues, which are very sensitive.”
Obstacles to the exchange of information hinder the work of doctors and research. To make the data accessible without compromising the privacy of patients, the AISym4MED project was developed.
“We are going to create data that is a representation of what is real and then we are going to change it a bit, enough so that we cannot recognize who originated this record,” the researcher tells TSF.
“A person can have an image of a skin problem in which a tattoo appears that reflects the logo of the person’s identity”, he exemplifies, and therefore the researchers commit to creating “a similar image, but removing the tattoo”, through artificial intelligence.
Thus, it is possible to guarantee that the identity of the person in question is secure. Researcher David Belo says that this technological advance allows decision-making to be “much better informed”, since in this way “medical science can finally have access to data that it did not have before”.
In addition, “it will be possible to make intelligent algorithms, with greater robustness and more prepared for the real world, that help doctors in their medical decisions.”
The project, which has funding of seven million euros, has been working with databases from hospitals in the Netherlands, Catalonia, the Basque Country and Turkey. The goal is to make information secure and accessible throughout the world.
Source: TSF
