Odds are you wouldn't DT. The changes are designed to be very subtle and gradually move the performance in the desired direction. As a pilot/rider you would not like to have the machine be markedly different every time the situation changes. The changes are usually stored in firmware so the machine starts off where it quit last time. Some of these controllers can be reset by pulling power off them like the Evo. Others will store changes semi-permanently until they are reflashed like many modern cars and trucks.In terms of Machine Learning algorithmic behaviour, I haven't detected that yet.
Nonetheless, learning and deep learning technology is making its way into everyday life.