It is following the path of least resistance. In this case, that path is the copper and the ions in the water that allow some more electrons. You need to look at water as a very high resistance. It isn't infinity, so some current will flow there, usually with bad results.

Another thing usually ignored is that ICs are not completely hermetically sealed, so water and other elements can go in and wreak havoc inside, especially with time when you take into account oxidation and ions that can short the IC.

It is following the path of least resistance. In this case, that path is the copper and the ions in the water that allow some more electrons. You need to look at water as a very high resistance. It isn't infinitely high, so some current will flow there, usually with bad results. Just as in the case with two resistors in parallel, the copper is the low resistance, and most current flows through that path, but some will flow in the other high resistor (ions in the water). Note that deionized water cannot conduct due to the lack of ions.

Another thing usually ignored is that ICs are not completely hermetically sealed, so water and other elements can go in and wreak havoc inside, especially with time when you take into account oxidation and ions that can short the IC.