Google’s ReCaptcha in version 3 works in the background. Instead of displaying images of crosswalks and such, it uses a kind of risk score. This risk score is based on user behavior: If someone has behaved like a human in the past and thus gets a low risk score, the captcha is passed without you having to do anything or even seeing it.
I assume that Google uses data from it’s own services, web analytics applications and usage data from Andoird devices and Chrome for this. Of course, this is not without its privacy issues but it’s convenient.
Yes, pretty much. But unlike Cloudflare DDos Protection and such Google ReCaptcha is mainly used to secure forms of all kinds (e. g. signup, login, contact or frontend posting forms).