LISA is an upcoming ESA mission that will detect gravitational waves in space by interferometrically measuring the separation between free-falling test masses at picometer precision. To reach the desired performance, LISA will employ the noise reduction technique time-delay interferometry (TDI), in which multiple raw interferometric readouts are time shifted and combined into the final scientific observables. Evaluating the performance in terms of these TDI variables requires careful tracking of how different noise sources propagate through TDI, as noise correlations might affect the performance in unexpected ways. One example of such potentially correlated noise is the relative intensity noise (RIN) of the six lasers aboard the three LISA satellites, which will couple into the interferometric phase measurements. In this article, we calculate the expected RIN levels based on the current mission architecture and the envisaged mitigation strategies. We find that strict requirements on the technical design reduce the effect from approximately 8.7 pm/rtHz per inter-spacecraft interferometer to that of a much lower sub-1 pm/rtHz noise, with typical characteristics of an uncorrelated readout noise after TDI. Our investigations underline the importance of sufficient balanced detection of the interferometric measurements.