In mice. This is a repeating trend in Alzheimer's research, where the amyloid-beta treatment works in the mouse model but not on humans, because the mouse model induces the amyloid-beta issue (mice don't really get Alzheimer's) and then we treat it.
Since mice don't ever get Alzheimer's naturally, and we don't actually know what Alzheimer's is, we don't know what it would even mean to give mice Alzheimer's. But for research we've genetically engineered mice that end up with lots of those plaques, and their behaviour does suggest an impairment similar to Alzheimer's, so that's what we've been working with. And in those models, various treatments that involve clearing the plaques does seem to help resolve that impairment - but they don't help humans with Alzheimer's, even if they do clear the plaques there too.
If I'm reading your question correctly, we can't stimulate amyloid plaque growth in humans for experimentation because that'd almost certainly be considered completely unethical. And also our methods for inducing the amyloid plaques involve mice that are genetically modified from birth rather than something we introduce in vivo, which would somehow be even more unethical than experimenting on live humans. It's possible we could make those genetic modifications in vivo now with recent developments in gene therapy, but...why?
Amyloid beta might not be causative, but if you hit a mechanism then it stands to reason it might be indicative - in this case if Alzheimer's is partly or fully caused by a waste removal problem in the brain.