Foreground galaxies that amplify the light from background quasars may also dim that light if the galaxies contain enough dust. Extinction by dust in lenses could hide the large number of lensed systems predicted for cosmologies with dominant cosmological constant. This could relax the strongest constraint on such cosmological scenarios, allowing a considerably older universe, as suggested by globular cluster ages. There is now increasing evidence for dust and gas in lensing galaxies, both from optical-near colors indicating reddening and from absorption by gas at the lens redshift. The dusty lens hypothesis makes two predictions that we propose to test with ISO observations. First, the strong 9.7 micron silicate feature should be visible in absorption at the lens redshift. Second, dust in the lensing galaxy should emit substantial energy at 60-200 microns. Our observations will constrain the incidence of dust in lenses. If dust is confirmed, we will also be able to study its detailed properties at high redshift, which is not often possible. We will observe with CAM CVF five lensed systems showing strong evidence for dust and gas. This will detect the silicate absorption feature for lensing galaxy dust up to redshift 0.7 and column density of A_V=1 magnitude if the Tau(silicate)/A_V ratio is similar to the Galactic one. For the nearest lens, Q2237+030 (lens redshift=0.039), we can distinguish the FIR bump of the lensing galaxy from the continuum of the quasar, which is at much higher redshift (z=1.69). This can be used to estimate the total amount of dust in the lensing galaxy.