Regression of differenced time series can lead to type I spurious regression, where the relation among their levels is rejected. This happens because differencing doesn't preserve the underlying relation due to systematic errors. Time series are dynamic processes, and ignoring their system dynamics can result in systematic errors in regression equations.