You are not understanding the trust issue.

By default, patents trust doctors with their bodies so they don't expect the doctor to do shady stuff like record them with a hidden camera. However, the moment they see something that suggest they might be recording,then the hidden voice awakens and questions everything the doctor has done in a new, not so trustworthy way.

It's not that patients think that doctors who wear those glasses won't have a hidden camera too, it's that doctors who have those glasses might also have a hidden camera, it's the fact that they have those glasses that opens the patient to other shady possibilities.

As others have said, trust is broken.

Yes, I totally understand the logic. It's just a transitivity problem, and can be modeled using a cost-benefit analysis:

1. A hidden camera is more dangerous for the victim because it is less risky for the creep/abuser.
2. You will not reasonably believe that a doctor will have a hidden camera inside his office.
3. A camera exposed to the naked eye is less dangerous for the victim because it is more risky for the creep/abuser.
   • The abuser is unlikely to take that risk because he could lose too much, much more than he would gain.
4. So it is reasonable not to believe that your doctor will record or photograph you with a camera exposed to the naked eye since it is inherently riskier for the creep, and it is not reasonable to believe otherwise.

Although, the problem here then shifts to the hardware itself. Like, since these lenses were probably closed source and proprietary hardware, and their software as well, then you don't have a way to prove that if you turn off the camera on said lens (either via the software kill-switch or via the hardware kill-switch) it is actually turning off, or just appears to turn off when it is actually still on but disguised as being turned off. And taking into account that there are precedents of devices doing this (a particularly famous case was that of a smart TV that even when you turned it off, it was not really off, but in a false shutdown, a very low power state where the screen was off, but the core and some telemetry processes, and the microphone, were still running and, thus, recording and sending everything that was recorded to their servers). So the only rational position, at least from a Game Theory point of view, is to assume that every device that is both closed source and closed hardware, that is not Free, Libre and Open Source Software, and Libre and Open Source and Schematics Hardware (yeah, even releasing their schematics), does the same thing.