I think that's taking it one step too far Eyewitness descriptions are, and have always been, one of the three major ways zoologists are led to new animals. There are really ONLY three ways (countless variations, but three main categories of events) in which a new animal CAN be discovered by science. They are 1) discovery of body parts (bones, trophies, things made from the animal's skin, etc); 2) scientific surveys where scientists are in the field looking for every animal in a targeted area; and 3) eyewitness accounts (either fresh or traditional) that alert scientists or explorers to the possibility of an animal and inspire expeditions to find it. Most of Dr. Alan Rabinowitz's mammal discoveries, for example, came from asking local hunters about their animals. Sometimes they could show him a trophy: other times, they described an animal they had seen and told him, or guided him, to where it could be found. As I said, there are many variations on these categories, but the idea that eyewitness encounters have not been crucial to important animal discoveries is certainly not valid.
(By the way, if you are curious what Dr. Rabinowitz has been up to, he's in the fight of his life: battling cancer while trying to save the tigers of Myanmar. I'm in awe of the man.)
Such sightings serve as a starting point for investigators: they are not "proof" of a creature, but they can prompt us to ask interesting questions which we can then approach with the modern tools of science. The sightings of the chevron-marked beaked whale called Mesoplodon Species A are a good example, leading eventually to an identification (which frankly still seems a little weak to me, but I have to yield to experts like Robert Pitman and company here) of this animal as the adult form of the pygmy beaked whale. Cryptozoology, properly understood, is the application of zoology, scientifically and objectively, to the discovery of new animals: the distinction is that cryptozoology opens the aperture a bit to open files on cases which are not quite as well attested as those leading to, say, the finding of the Vu Quang ox and company
What is the eyewitness report is not followed by anything more substantial? At what point do we toss it out?
Let's say it's 1908 or so, and you open a sea serpent file based on the report made by two naturalists on the yacht Valhalla. Interesting sighting, just published in the Royal Society's Proceedings - perfectly logical thing for a scientist to do. Then you wait. Do you close the file if twenty years pass without the animal being found? Probably not - the sea is a big place. Fifty years? Maybe - 50 years without a sighting was the old IUCN standard for extinction. 100 years? Well, depressingly, it's entirely logical to close the file. (I haven't quite, but I recognize I'm on shaky ground). In other words, how long does it take for absence of evidence to become evidence of absence? Maybe there should be a 50-year standard, but the cahow or Bermuda petrel was rediscovered 300 years after extinction. Some of it depends on whether the habitat can be searched: small lakes have been thoroughly searched (and dynamited) and the hypothesis (in Karl Popper’s sense of the falsifiable hypothesis being the basis of science) that there were creatures in those lakes have been properly falsified. It would take enormous and unavailable resources to falsify the hypothesis "There is an unclassified North American ape," but you can do it in theory. For the hypothesis, "There is an unclassified elongate marine species sometimes called a sea serpent" you could still falsify it in theory by active searching, but the task is too vast to even consider. Can the lack of followup evidence be considered falsification, and after what period of time? You inevitably end up in the world of opinion.
It’s not true that “my opinion is as good as yours” (see the Pitman example above). But it’s also true that every researcher needs to use their own judgment – hopefully skeptically (in the proper sense of that word) – when evaluating witness reports. Witnesses can be right, they can be wrong, or somewhere in the middle. But I do hold they very often give science the starting point in discovery of a new animal.