The same is not true of a series of images on a screen in an isolated room. Images on screens are not people. There is no social conditioning that normal human beings have that will cause people to automatically respect images on a screen.
I'd argue that has pretty much been proven by the Milgram experiment.
While I believe the article is correct, and later experiments probably bear it out, Milgram's experiments don't directly support that cliam: The "teacher" (the subject) could not see the person he believed he was shocking at all.
Nonetheless, it's a relevant link to bring up because of the chain of command and obedience to authority aspects at play here.
In another variation, the teacher actually had to physically press the (fake) pupil's hands onto the device that gave electric shocks. IIRC, this squicked out a few more people, but quite a large portion still went through with it.
Proximity to the teacher was the biggest influencer rather than proximity to the student. When the teacher was further away (in some simply on a phone) the subject would more frequently refuse, more so than when they were in close proximity to the student.
Proximity to the student shouldn't be ruled out, after all it would get a few more people behaving like they really should, however its these screening people's proximity to their bosses that is the biggest influence on whether they treat a person with humanity or not.
I'd argue that has pretty much been proven by the Milgram experiment.
http://en.wikipedia.org/wiki/Milgram_experiment