This is true. It's like optical illusions, such as
lines appearing to be different lengths, even though they're the same. Most people "see" the illusion, though it varies among cultures.
Like pareidolia, there are two "truths" involved, each backed by different kinds of evidence. We can estimate the length of the lines by comparing them to each other by eye, or we can use a ruler to measure the length, and each yields a different answer.
Measuring the length doesn't stop the optical illusion (at least it doesn't for me--I still see the lines as different lengths if I take the ruler away), but like most people, I trust the evidence of the ruler instead of the estimate by eye.
Imagine a person who refused to believe the ruler measurement, and insisted that because the lines looked to be different lengths, they actually
were different lengths, and came up with explanations for why that was so. Perhaps the computer screen could sense when a ruler was near and changed the lines so they'd be the same length for that moment. Perhaps every ruler in the house had been secretly replaced by special ones that read wrong.
The explanations would sound ridiculous to most people, who would be mildly surprised their eyes had been fooled but would have no emotional investment in the length of the lines and would immediately accept the ruler measurement as the true one, based on the best evidence.
Pareidolia is like the optical illusion that most people see. An objective test, such as the one for the Million Dollar Challenge, would be like laying a ruler along the lines. If the test fails to show that the voices are anything more than the effects of pareidolia, that's where the conclusions of skeptics and believers would diverge. Believers would have so much emotionally invested in the perception caused by pareidolia, they'd search for any other explanation. Skeptics would accept that there was now stronger evidence the voices were "real," and proceed from there, with further tests to confirm it and new hypothesis about what's happening.
Right now, though, skeptics aren't predicting that conclusion, because it would be like someone showing two lines that seem to be different lengths and claiming there was no illusion this time. Every other time it's been investigated, it's turned out to be an optical illusion, so there's no reason to predict the results will be different when we lay the ruler down this time.
You've been blaming people's skepticism about the voices on fear, but I think the fear is stronger the other way. The potential of having to abandon a belief that's emotionally important is frightening.