Dark stars are luminous stellar objects, made primarily of hydrogen and helium, yet powered by dark matter. Formed in the early Universe at redshift z =10-20, some can grow supermassive by accretion. We investigate the detectability of supermassive dark stars (SMDSs) by the Roman Space Telescope (RST), motivated by our recent identification of several SMDSs candidates in the James Webb Space Telescope (JWST) data. Similarly to JWST, RST will be able to detect SMDSs at redshifts as high as z = 14 as photometric dropouts. Thanks to its much larger effective field of view, RST will have a larger detection probability of SMDSs candidates at these redshifts (compared to JWST). JWST, on the other hand, being sensitive to higher wavelengths, can probe a larger swath of the reframe SEDs of any given high redshift objects, and, as such, confirm the Dark Star nature of RST candidates via spectroscopy with NIRSpec or photometry with MIRI/NIRCam. In the talk I will elaborate on the various possible ways to differentiate between SMDSs and early galaxies using photometry, image morphology, and spectroscopy. The detection and confirmation of any supermassive dark star will provide evidence for annihilating dark matter. Moreover, such massive stars can also be natural progenitors of the supermassive black holes powering the extremely bright quasars observed at early times.