It's tough to be clear about the risks of these devices, since there is very little public information on their nature.
Both THz and X-ray scanners have been proposed to look for hidden weapons and explosives, with X-ray machines detecting the enhanced scattering of X-rays compared with flesh, and THz machines detecting the different temperature profiles of a hidden object and flesh.
The machines being used in the US do seem to be based on X-ray scattering. I was initially astonished by this, since unnecessary exposure to ionizing radiation is not usually promoted. Whether the employment of the former homeland security secretary Michael Chertoff by an X-ray machine manufacturer has any relevance to this decision is an interesting question.
From looking at quoted numbers previously, and a rescan of wikipedia
, I think that the `Letter of Concern' (unsurprisingly) contains very valid points. The claimed dose from a scan is comparable to the background radiation dose at sea level in an hour, and much less than the enhanced cosmic-ray dose from a 1-hr flight at 30000ft; however, the energy spectrum and directionality of these sources is different, and so any health effects may be too.
Having a scan before each 5-hour flight should not significantly increase a pilot's total radiation exposure, but why not opt for the non-ionizing patdown instead?
The letter raises the particularly valid concern that it's easier to see how the machines would overdose than underdose. If they were under-powered, the images would probably not be useful and the manufacturer would be called straightaway (I assume that the absolute minimum dose required to obtain a useful image is used). However, the operators would not know to call the manufacturer if the thing was going wild, unless the detectors were saturating. I would hope there are self-diagnostic checks built in, and the absolute dose is frequently measured.