You actually can, and people routinely do. You just need to run the experiment past an IRB first. Which, incidentally, Arday didn't do for the studies he claimed to have done. Plus, laboratory studies aren't the only way to study people or any other animal. Observational studies happen all the time, and when well designed they can be incredibly useful. But autoethnographies are not valid observational studies.
It's a useful tool for inflating your publication count, sure. That's why it's gained such popularity. And cross referencing other works doesn't mean jack ◊◊◊◊ if the underlying work you're doing is worthless.
Research should have rigor. Autoethnographies do not.
No one objects to researching growing up in Toxteth. But recounting your own personal life isn't a rigorous way to do that. If you actually want to research growing up on Toxteth, you could do things like study income distributions in the area, look at educational achievements of students from the local schools, etc. And you could do all of that with rigor too.
Arday didn't do any of that.