So the only conceivable way this works on all types of hair is by using NIR lasers.
A couple Watt of <1300nm NIR is not eye safe though. The neat concept of FTIR would help with detecting hair to save on power but it can't contain the light. A simple water drop could couple out the light as well and if it can burn a hair it can burn a spot on your retina.
Large amounts of the classical 1970s Laser Safety research are starting to show as being very conservative. Things like audience scanning during laser light shows in Europe and recent tests involving exposure to laser pointers have shown the damage thresholds for retinas that were put into place may be extra conservative. I'm a certified LSO for laser shows, my old hobby. So I keep an "eye" on this area. I'm usually the de-facto LSO for the area I'm working in, as not many organizations these days can afford a full time LSO. Remember when the standards were formed in the 1970s, calibrated devices for measuring laser power were in their infancy.
Couple that with high multiplication factors tacked on by researchers who
RIGHTFULLY wanted to ensure public safety, and you can see that some of the numbers are very high. Back then they did not have confocal imaging to see lesions or photochemical damage on the test animal tissue, either.
I still tell people never to fail to take precautions around with anything hotter then Class IIIA devices, but some of the classic thresholds and teachings are very, very conservative. There is a current campaign to allow "Ten Times MPE" for audience scanned laser shows by trained operators with certified fail safes in the scanning gear.
That said, the only way to ensure laser safety in ANY real world situation is to NEVER assume anything and take actual measurements.
Steve