It's pretty easy to test if people can detect being stared at, and yet no one has bothered to set up a clean demonstration. Why not?
This is far removed from "prey detecting whether a predator is staring at them", which would be a much harder test to set up.
For humans, "prey" might mean someone with genuine sexual interest in you, staring at your backside. Try setting that up as a scientific test! Observer effect in full swing, and your results will be disappointing.
Not that I believe or support ideas because they're exotic and exciting. My point here is that it's worth keeping an open mind, precisely because science is entering uncomfortable areas such as consciousness, ego, identity, dreams.
It might be the case that the "sense of being stared at" is real but it could easily be a combination of ordinary senses. By isolating the confounding variables scientists can determine whether this is an independent effect.
It may be that the very act of staring at someone actually alters or influences something in the physical world, as those flowers do we've recently learned.... >"electrical fields—may work in concert with these other cues to provide extremely nuanced details about pollination status."
Along those lines, it's plausible that there's processes at work with staring we're currently unaware of, waiting for science to discover.
Better that fictional ideas about possible nature of matter come from scientists anyway. Having a go at guessing where science may uncover new mechanisms, or where current science deserves more attention, is a good thing to have a go at.
It's the equivalent of drawing attention to a new patch of ground for a fresh archeological dig, he's not actually claiming on the TEDx stage to have dug it up and found an alien artifact.