If SpaceX cared, the Starlink fleet (both satellites and ground stations) could be the biggest, most sensitive radio telescope in the world, without compromising its usefulness as a network. It points in all directions at the same time. With such sensitivity, a very short "exposure" time is plenty.
Each node records a few cycles of analog waveform at a certain atomic-clock time determined by its position, and forwards that to a common collection point to correlate with the others.
The Starlink satellite antennas all point at the Earth. The antennas are also small compared to radio telescopes. They also listen to specific frequencies. They would make the best radio telescope for detecting Starlink customers.
The customer antennas would make poor radio telescope because the antennas are small. If using array of small antennas as radio telescopes were possible, we would see them. Using interferometry requires measuring positions and times to sub-wavelength. The Square Kilometer Array is the closest, but it is small extent, with large dishes, and many low-frequency antennas.
Each node records a few cycles of analog waveform at a certain atomic-clock time determined by its position, and forwards that to a common collection point to correlate with the others.