Suppose the true relationship between two variables is $y = 2.0x + 5.0 + \epsilon$. However, the researcher cannot measure $x$ directly and instead measures $x^* = x + u$, where $u$ is a random measurement error that is independent of $x$ and $\epsilon$. If the researcher fits a simple linear regression of $y$ on $x^*$, what is the effect of this measurement error on the estimated slope?