In this case, I believe, the prevalence of cheating among CS students isn't higher than cheating among students of other courses in the same universities, or elsewhere. Even in public health, occasionally, the prevalence of a certain disease goes up courtesy of technological advancement. The prevalence of that disease in a population that was initially low or unknown suddenly shoots up because the means with which to detect the disease has now become widely available, and not always because the incidence of that disease in that population has only started going up. Now, if other courses offered at these named universities had means of detecting students cheating in such courses, the prevalence of cheating in these other courses would probably be as high as in CS. Universities would thus need to either come up with systems that detect cheaters that has both a high true positives rate and a low false negatives rate, or, come up with a better way of assessing students that does not involve exercises that can easily be "borrowed" from the Internet.