You can easily pick out the recording LED on camera equipment from 30m so that's not really an issue. The bigger problem is simply not noticing once glass becomes even slightly popular.
You could easily make it completely opensource and standalone by replacing the calls to web-based api's like lambda or betaface, with the standalone face recognition engine from openbiometrics.org https://github.com/biometrics/openbr
If they don't like it its only because they want to be the only ones doing it. Seriously facial recognition is the one big technology useful on the glass and Google knows that and probably developed a lot for the glass itself.
It doesn't use lambda labs since they weren't out yet, but it is just as easy to use other APIs anyway.