Couldn't they just use an image intensifier tube for night vision? I think they should just stick with the excuse of it being costly and unnecessary.
So in this sense webb is already a camera which can photograph itself!
This capability exists for mirror alignment: They'll point webb at an isolated star, switch to focusing on the primary mirror, swap in optics that cause small phase differences to result in diffraction patterns, and then they can use the resulting images to fine tune the positioning of the mirror segments to a small fraction of the wavelength of light that they're using.
I'm not sure if that's correct - there's a lot of photons hitting the telescope (the night sky), and night vision systems can work off of a relatively small number of photons