The paper summarized at the link at the top hypothesizes that in a setup like this:
Bb --- Earth --- molecular cloud --- B'b'
where Bb and B'b' are very similar (from the molecular cloud's perspective) inspiralling binary black hole pairs, and Earth is where LIGO (and Virgo and other detectors) are, then we will see characteristic bright flashes from within the molecular cloud and a change in the B'b' detected waveform because it will have lost some energy to the production of the flashes of light.
(The light flash will trail behind the dampened B'b' waveform detection because this hypothesized mechanism only works when a refractive medium slows light from c (its speed in vacuum) but does not slow gravitational waves from c. Although the authors do not touch on the matter, I suspect that we would also be interested in <https://en.wikipedia.org/wiki/Light_echo>s.)
The paper focuses on an analysis of water as the dominant molecule. There are known astrophysical water megamasers (see second paragraph at <https://en.wikipedia.org/wiki/Megamaser>), so this is far from ridiculous.
