The implication is that the users that are being constantly presented with CAPTCHAs are experiencing that because they are unwittingly proxying scrapers through their devices via malicious apps they've installed.
or just that they don't run windows/mac OS with chome like everyone else and it's "suspicious".
I get cloudflare capchas all the time with firefox on linux... (and I'm pretty sure there's no such app in my home network!)
When a random device on your network gets infected with crap like this, your network becomes a bot egress point, and anti bot networks respond appropriately. Cloudflare, Akamai, even Google will start showing CAPTCHAs for every website they protect when your network starts hitting random servers with scrapers or DDoS attacks.
This is even worse with CG-NAT if you don't have IPv6 to solve the CG-NAT problem.
I don't think the data they collect is used to train anything these days. Cloudflare is using AI generated images for CAPTCHAs and Google's actual CAPTCHAs are easier for bots than humans at this point (it's the passive monitoring that makes it still work a little bit).