[Artificial intelligence applied to the electrocardiogram, or is there really a needle in a haystack?]

Artificial intelligence applied to the standard Ecg (Ai-Ecg) is able to enormously enhance the performance of a diagnostic test that has been in use for over 100 years. The tool can in fact identify diseases, cardiac and non-cardiac, hidden or foreseeable, often asymptomatic, previously undetectable. Here we describe some of the main applications in cardiology: sub-clinical atrial fibrillation and asymptomatic left ventricular dysfunction. However, the Ai-Ecg still needs validation in non-experimental contexts: only in that case can its potential be considered verified and therefore usable in clinical practice.